IMPACTS AND MANAGEMENT OF CENCHRUS CILIARIS (BUFFEL … · 3.2.3. Fire as a management tool for...

IMPACTS AND MANAGEMENT OF CENCHRUS CILIARIS (BUFFEL GRASS) AS AN INVASIVE SPECIES IN NORTHERN

QUEENSLAND

Thesis submitted by Janice JACKSON B.Agr.Sc. (Hons 1) (U.Q)

in May 2004

For the degree of Doctor of Philosophy In Tropical Plant Sciences School of Tropical Biology

James Cook University

STATEMENT OF ACCESS I, the undersigned, the author of this thesis, understand that James Cook University will make it available for use within the University Library and, by microfilm or other means, allow access to users in other approved libraries. All users consulting this thesis will have to sign the following statement:

In consulting this thesis I agree not to copy or closely paraphrase it in whole or in part without the written consent of the author; and to make proper public written acknowledgement for any assistance which I have obtained from it.

…………………………………………… …………………………… Janice Jackson STATEMENT OF SOURCES DECLARATION I declare that this thesis is my own work and has not been submitted in any form for another degree or diploma at any university or other institution of tertiary education. Information derived from the published or unpublished work of others has been acknowledged in the text and a list of references is given. …………………………………………… …………………………… Janice Jackson

ii

Abstract

Cenchrus ciliaris L. (buffel grass) (Poaceae) is recognized as one of Australia’s most serious

environmental weeds. This introduced grass is associated with loss of native species and

alteration of fire regimes. However there is considerable controversy regarding its weed

status as it is also highly valued as a pasture species for arid and semi-arid zones.

Quantitative studies are needed to determine its ecological effects. In addition, its spread into

non-target areas, including conservation reserves, means that there is considerable interest in

strategies for containing or eliminating C. ciliaris. These two issues, the effects of C. ciliaris

on native species and strategies for managing C. ciliaris, are the focus of this thesis.

The relationship between C. ciliaris and herbaceous species richness was investigated in two

studies at a range of scales up to 64 m2 in open woodlands in north-eastern Queensland. In

the first study, the herbaceous species composition of sites with and without C. ciliaris were

compared. Cenchrus ciliaris-dominated sites had fewer herbaceous species than non-C.

ciliaris sites at all scales investigated and this pattern was found for the major plant groups

(perennial grasses, legumes and other forbs) present. In the second study, the relationship

between varying levels of C. ciliaris biomass and species richness was investigated at one

site. The relationship between varying levels of a dominant native grass, Bothriochloa

ewartiana (Domin) C.E. Hubb. (Poaceae), and species richness was also determined for

comparison with the C. ciliaris biomass-richness relationship. In this study, species richness

was negatively associated with increasing C. ciliaris biomass at some scales and it appeared

that C. ciliaris had a greater effect on richness than B. ewartiana. The negative association

between C. ciliaris and species richness is consistent with the view that invasion by C.

ciliaris poses a threat to biodiversity. However, the precise cause of the relationship has yet

to be determined.

The strategic use of fire offers potential to control unwanted species. To evaluate fire as a

tool for reducing C. ciliaris abundance, the effects of season of burning on two C. ciliaris-

dominated communities in north Queensland were investigated. Three treatments were

imposed in small plots at both sites: early dry season burn, late dry season burn and control

(no burn). These treatments were selected to exploit differences in fire characteristics and

vegetation responses to fire associated with different season of burning. The herbaceous

species present and their cover were recorded before and after the fires and post-fire seedling

emergence was monitored. To help understand the mechanisms by which fire may alter

community composition, burning treatment effects on the availability of establishment sites

and propagules were also investigated. Fire affects establishment site availability by

iii

reducing resident plant competition, by altering nutrient availability and by altering soil

surface condition. Three studies were conducted to investigate treatment effects on

establishment sites: (1) C. ciliaris plants were monitored to determine mortality, (2) a

bioassay technique was used to assess plant nutrient availability and (3) a ‘pot’ experiment

was conducted to examine the effects of different soil surface cover on seedling emergence

to help predict the effects of litter removal on emergence patterns. Fire effects on propagule

supply were investigated by monitoring flowering in C. ciliaris. A germination method was

used to determine soil seed bank composition.

Overall, burning had little effect on these communities. The intensities of the fires were low

to moderate (300-3030 kWm-1). At Dalrymple there was an unexpected reversal of

intensities; the mean intensity of early dry season fires was higher than that of late dry

season fires. The fires caused no major changes in composition, few C. ciliaris plants were

killed and no changes in nutrient availability or seed bank composition were detected.

Although these short-term studies of single fires do not allow definitive recommendations

regarding the use of fire to manage C. ciliaris, they provide information that will aid future

research. I found that fire could kill C. ciliaris plants and reduce C. ciliaris cover. This

contrasts with the positive fire feedback model generally proposed for C. ciliaris. Cenchrus

ciliaris mortality was higher with early dry season burning at Dalrymple, suggesting that

higher intensity fires will be more effective in eliminating C. ciliaris plants and/or that C.

ciliaris plants may be more susceptible to fire at this time because they have not fully

senesced. Apparent low densities of perennial grass seeds in the seed banks of these

communities may be exploited: over-sowing with native perennial grasses after fire may

encourage shifts in perennial grass dominance.

There is an urgent need for management strategies that reduce, prevent or contain invasive

weed invasion. Further work is required to investigate the application of fire regimes in C.

ciliaris-dominated communities. Of particular interest are differences in growth and/or

phenology between C. ciliaris and native species in these grasslands that may be exploited to

disadvantage C. ciliaris.

iv

CONTENTS

Statement of Sources Declaration………………………….…….……………….…...……ii

Statement of Access………………………………………….…….………………..……….ii

Abstract………………………………………………………………….…………..………iii

List of Tables…………………………………….…………………….…..………...………ix

List of Figures……………………..……………………………………………...….………x

Acknowledgements………………………………………………………………....……...xiv

CHAPTER 1. GENERAL INTRODUCTION 1

1.1. Cenchrus ciliaris (buffel grass) – wonder grass or weed? 1

1.2. The Issues: impacts and management 2

1.3. This project 3

CHAPTER 2. CENCHRUS CILIARIS AS AN INVASIVE SPECIES 5

2.1. Literature review: impacts of invasive plants 5

2.1.1. Introduction 5

2.1.1.1. This review 6

2.1.1.2. A note on terminology 6

2.1.2. Biological invasions 8

2.1.3. Impacts of invasive plants 10

2.1.3.1. Impacts on ecosystem structure/components 11

2.1.3.2. Impacts on ecosystem function/processes 15

2.1.3.3. Mechanisms underlying invasive plant impacts 21

2.1.4. Grasses: a family of problem plants 22

2.1.5. Invasive plants in Australia 25

2.1.5.1. Weeds or wonder plants 27

2.1.6. Cenchrus ciliaris 29

2.1.6.1. Origin, introduction and establishment in Australia 30

2.1.6.2. Physical description and growth characteristics 31

2.1.6.3. Cenchrus ciliaris – valuable pasture species or destructive invader? 32

2.1.7. Measurement and assessment of invasive plant impacts 34

2.1.7.1. What to measure 34

2.1.7.2. How to measure plant impacts 35

2.1.7.3. Impact assessment 36

v

2.1.8. Conclusions 37

2.2. Is there a relationship between herbaceous species richness and Cenchrus

ciliaris abundance? 39


2.2.2. Methods 40

2.2.2.1. Study area 40

2.2.2.2. Study one: herbaceous species richness with and without Cenchrus

ciliaris

41

2.2.2.3. Study two: herbaceous species richness with varying Cenchrus

ciliaris biomass

42

2.2.3. Results 45

2.2.3.1. Study one: herbaceous species richness with and without Cenchrus

ciliaris

45

2.2.3.2. Study two: herbaceous species richness with varying Cenchrus

ciliaris biomass

48

2.2.4. Discussion 53

CHAPTER 3. MANAGING CENCHRUS CILIARIS-DOMINATED

VEGETATION WITH FIRE 60

3.1. Introduction 60

3.2. Literature review: the role of fire as a vegetation management tool 64


3.2.2. Fire: an ecological phenomenon 65

3.2.2.1. Fire type 65

3.2.2.2. Fire behaviour 66

3.2.2.3. Fire regime 67

3.2.2.4. The effects of fire 69

3.2.2.5. Plant responses to fire 70

3.2.3. Fire as a management tool for invasive plant control 73

3.2.3.1. The importance of fire regime 74

3.2.3.2. The importance of vegetation characteristics 77

3.2.3.3. The importance of other factors 80

3.2.3.4. The efficacy of fire as a tool to manage vegetation composition 81

3.2.3.5. Community responses to fire: creating gaps and filling them 84

3.2.4. Cenchrus ciliaris and fire 85

3.3. Site and treatment descriptions 88

vi

3.3.1. Study sites 88

3.3.1.1. Dalrymple National Park 88

3.3.1.2. Moorrinya National Park 90

3.3.2. Experimental design and treatments 90

3.3.3. Pre-fire measurements 91

3.3.3.1. Plant species composition 91

3.3.3.2. Fuel biomass and moisture content 92

3.3.3.3. Weather conditions 92

3.3.4. Fire conditions 93

3.3.4.1. Early dry season fires 93

3.3.4.2. Late dry season fires 94

3.3.5. Local rainfall 95

3.4. Can fire kill Cenchrus ciliaris? 96

3.4.1. Introductions 96

3.4.2. Methods 97

3.4.2.1. Field set-up 97

3.4.2.2. Statistical analyses 98

3.4.3. Results 98


3.5. Does fire increase plant nutrient availability in Cenchrus ciliaris-dominated

grassland? 104


3.5.2. Methods 106

3.5.2.1. Soil collection and processing 106

3.5.2.2. Pot set up 107

3.5.2.3. Statistical analysis 108

3.5.3. Results 108


3.6. Litter: a help or a hindrance to seedling emergence? 112


3.6.2. Methods 116

3.6.2.1. Treatments 116

3.6.2.2. Experimental details 116


3.6.3. Results 119


vii

3.7. Effects of season of burning on soil seed banks of Cenchrus ciliaris-dominated

grassland 129


3.7.2. Methods 131

3.7.2.1. Soil collection and processing 131

3.7.2.2. Glasshouse set up 131

3.7.2.3. Measurements 132


3.7.3. Results 133


3.8. Does fire promote flowering in Cenchrus ciliaris? 144


3.8.2. Methods 144

3.8.2.1. Measurements 144


3.8.3. Results 145


3.9. Effects of season of burning on seedling emergence patterns in Cenchrus

ciliaris-dominated grassland 149


3.9.2. Methods 151

3.9.2.1. Seedling monitoring sites 151

3.9.2.2. Study 1: seedling recruitment study 151

3.9.2.3. Study 2: grass seedling recruitment and survival 151

3.9.3. Results 153


3.10. Effects of season of burning on herbaceous community composition of

Cenchrus ciliaris-dominated grassland 160


3.10.2. Methods 161

3.10.2.1. Plant surveys 161


3.10.3. Results 163


3.11. General discussion: managing Cenchrus ciliaris with fire 174

3.11.1. The effects of season of burning on Cenchrus ciliaris-dominated

viii

grassland 174

3.11.1.1. Effects of fire on establishment sites 175

3.11.1.2. Effects of fire on propagule supply 177

3.11.2. Fire as a management tool – are there opportunities to reduce Cenchrus

ciliaris abundance?

179

3.11.3. Conclusions 181

CHAPTER 4. CENCHRUS CILIARIS AS AN INVASIVE SPECIES – FUTURE

RESEARCH QUESTIONS 183

REFERENCES 186

APPENDICES 215

1A. Herbaceous species found in C. ciliaris and non-C. ciliaris plots in the

Dalrymple Shire survey

215

1B. Herbaceous species found in surveyed plots at Hillgrove 218

2A. Aussie peat components 219

2B. Germinable seed content of Cenchrus ciliaris and Heteropogon contortus

seed material

219

3A. Herbaceous species found in the Dalrymple seed banks 220

3B. Herbaceous species found in the Moorrinya seed banks 221

4A. Herbaceous species found in burning treatment plots at Dalrymple 222

4B. Herbaceous species found in burning treatment plots at Moorrinya 223

LIST OF TABLES

Table 2.1. Some introduced grasses considered to be problematic in Australia. 27

Table 2.2. Latitude and longitude, soil type and current grazing regimes of C.

ciliaris and non-C. ciliaris plots surveyed in the Dalrymple Shire.

41

Table 2.3. Summary of regression analyses investigating relationships between

biomass and herbaceous species richness at scales 2-64 m2 for (a) C.

ciliaris and (b) B. ewartiana.

51

Table 2.4. Summary of regression analyses investigating relationships between

ix

C. ciliaris biomass and the numbers of species of forbs, legumes, forbs

plus legumes and grasses at scales from 2-64 m2.

52

Table 3.1. Summary of conditions for early and late dry season fires at

Dalrymple and Moorrinya National Parks.

94

Table 3.2. Mean basal area (cm2) and mortality (%) of tagged plants at Dalrymple

and Moorrinya.

99

LIST OF FIGURES

Figure 2.1. Mean (± SE) numbers of species for C. ciliaris (■) and non-C. ciliaris

plots (□) at each of seven scales from 1-64 m2. Figure (a) shows the

total number of species. Figures (b-d) show numbers of non-

leguminous forb, legume and perennial grass species respectively.

46

Figure 2.2. Proportional composition, in terms of numbers of non-leguminous

forb, legume, sedge, annual grass and perennial grass species, of C.

ciliaris plots (■) and non-C. ciliaris plots (□) at (a) 1 m2 and (b) 64

m2 scales.

47

Figure 2.3. Herbaceous taxa showing contrasting distribution between C. ciliaris

and non-C. ciliaris plots. The number of shaded cells indicates the

number of plots in which the species was present.

48

Figure 2.4. Composition of 18 plots in terms of C. ciliaris ( ), B. ewartiana ( ),

and other herbaceous species ( ) biomass (g/m2) at the 64 m2 scale,

presented in order of declining C. ciliaris biomass.

49

Figure 2.5. Herbaceous species number in relation to dominant species (C. ciliaris

or B. ewartiana) biomass (g/m2) for C. ciliaris ( ) and B. ewartiana

(□) dominated plots at (a) 0.25 m2, (b) 0.5 m2 and (c) 1 m2 scales.

50

Figure 3.1. Location of Dalrymple and Moorrinya National Parks. 89

Figure 3.2. Monthly precipitation (bars) recorded from June 1998 to June 2001 at

(a) Fletcher View Station near Dalrymple National Park and (b)

x

Moorrinya National Park (except from September 1999 to September

2000 when precipitation for nearby Uanda Station is given).

95

Figure 3.3. Percentage of tagged plants in basal area size classes (increments of

100 cm2). (a) C. ciliaris plants at Dalrymple, (b) C. ciliaris plants at

Moorrinya and (c) Astrebla plants at Moorrinya.

100

Figure 3.4. Relationships between plot fire intensity and percentage of tagged

plants that died/plot at (♦) Dalrymple and (□) Moorrinya.

101

Figure 3.5. Mean (± SE) total above-ground biomass (g/pot) of sorghum plants

grown in soil from ( ) control (no burn), ( ) early dry season burn

and ( ) late dry season burn plots at Dalrymple and Moorrinya.

108

Figure 3.6. Relationship between plot fire intensity (kW/m) and mean total above

ground sorghum biomass (g/pot) for (♦) Dalrymple and ( )

Moorrinya.

109

Figure 3.7 Litter mat between C. ciliaris tussocks at Dalrymple 115

Figure 3.8. Soil surface cover types with H. contortus seed. From left to right:

eucalypt litter, matted litter, bare soil and open litter.

118

Figure 3.9. Mean percentage emergence of (a) C. ciliaris seedlings and (b) H.

contortus seedlings from four cover types (bare soil (♦), matted litter

(∆), open litter (X) and eucalypt litter (□)) over 21 days.

119

Figure 3.10. Mean percentage emergence of C. ciliaris plus H. contortus seedlings

from four cover types (bare soil (♦), matted litter (∆), open litter (X) and

eucalypt litter (□)) from seed sown top of litter (solid line) and under

litter (dashed line) over 21 days.

120

Figure 3.11. Mean percentage emergence of C. ciliaris seedlings (X), and H.

contortus seedlings (□) from seed sown on top of litter (solid line) and

under litter (dashed line) over 21 days.

121

Figure 3.12. Mean days taken for 50% emergence of C. ciliaris seedlings (lower

solid bars) and H. contortus seedlings (upper unfilled bars) from seed

sown on bare soil and under and on top of three litter types.

122

Figure 3.13. Mean (± SE) percentage of soil cover types at Dalrymple in January 122

xi

2001.

Figure 3.14. Mean number (counts/m2) of (a) annual grass, (b) perennial grass, (c)

forb, (d) legume and (e) sedge seedlings from unburnt (control), early

dry season and late dry season burnt seed banks from Dalrymple (1st

column) and Moorrinya (2nd column).

134

Figure 3.15. Percentages of perennial grass, annual grass, forb, legume and sedge

plant groups making up the seed bank seedlings (1st bar) and post-fire

standing herbaceous plant cover (2nd bar) at Dalrymple (left) and

Moorrinya (right).

135

Figure 3.16. Mean species richness (number of species/plot) of seedlings emerging

from unburnt (control), early dry season and late dry season burnt seed

banks from Dalrymple and Moorrinya.

136

Figure 3.17. Relationships between plot fire intensity (kW/m) and plot species

richness (number of species/plot) for (a) Dalrymple and (b) Moorrinya

seed banks.

136

Figure 3.18. Mean number of inflorescences per tagged plant at (a) Dalrymple (all

plants in January 2000) and (b and c) Moorrinya (C. ciliaris and

Astrebla spp. plants respectively in February 2000) in control (no

burn), early dry season fire and late dry season fire treatments.

146

Figure 3.19. Mean percentage of tagged plants in flower at (a) Dalrymple (all

plants in January 2000) and (b and c) Moorrinya (C. ciliaris and

Astrebla spp. plants respectively in February 2000) in control (no

burn), early dry season fire and late dry season fire treatments.

147

Figure 3.20. Mean number of forb seedlings (solid bars) and grass seedlings

(unfilled bars) in control, early dry season and late dry season fire

plots at Dalrymple in December 1999. Different lower case letters

denote significantly different forb seedling numbers.

153

Figure 3.21. Relationships between estimated plot fire intensity and mean number

of grass (∆) and forb (■) seedlings/m2.

154

Figure 3.22. Mean number of grass seedlings (seedlings/m2) found in control (no

burn), early dry season and late dry season fire treatments showing

seedling status (dead ■ and surviving □) at the end of the 2000-2001

xii

growing season (June 2001). 155

Figure 3.23. Mean (± SE) species richness (number of species/m2) in control, early

and late dry season burn treatments for (a,b) annual grasses, (c,d)

perennial grasses, (e,f) non-leguminous forbs and (g,h) legumes at

Dalrymple (1st column) and Moorrinya (2nd column).

164

Figure 3.24. Mean C. ciliaris percentage cover (shaded portion) and total perennial

grass percentage cover (bar) cover at (a) Dalrymple and (b)

Moorrinya.

165

Figure 3.25 Mean (± SE) percentage cover of (a) annual grasses, (b) forbs, (c)

legumes and (d) sedges at Dalrymple (1st column) and Moorrinya (2nd

column).

166

Figure 3.26 Relationship between fire intensity and perennial grass cover at (∆)

Dalrymple and (■) Moorrinya.

167

xiii

ACKNOWLEDGEMENTS This project would not have been possible without the support of Paul Williams, Queensland

Parks and Wildlife Service. Paul provided access to field sites and assistance with field work

and I thank him very much for his support and enthusiasm. Thanks also to QPWS staff

Benton Haigh, Ken McMahon and Eddie Staier who assisted with the fire work and others

who assisted with access to field sites and with field work including Brett Abbott, Peter

Allen, Jeff Corfield, Mandy Kotzman, Brigid McCallum, Mike Nicholas, Peter O’Reagain,

Ian Radford, Gary Rogers, Lynn Walker, Lindsay Whiteman and Mike Whiting. John

Childs, while director of the Tropical Savanna CRC, approved funding to support some of

this work. I would also like to thank him for financial support to attend a conference in

Townsville and a writing course in Darwin. Thanks also to Brett Abbott for regular help with

computer matters, Christy Matthes for library assistance and Chris Stokes for informative

discussions regarding species area curves among other things. Many thanks to Betsy Jackes

for seeing me through the university hoops. She and Tony Grice provided valuable feed back

during the writing of this thesis. I would also like to thank John McIvor, John Ludwig and

other reviewers for their advice on an associated manuscript that assisted with the writing of

the second chapter of this thesis. The various studies were undertaken in consultation with a

very patient statistician. Thank you Bob Mayer, senior biometrician with the Department of

Primary Industries. Finally, I would like to thank friends, family and my husband for their

support.

xiv

CHAPTER 1. GENERAL INTRODUCTION

1.1. CENCHRUS CILIARIS (BUFFEL GRASS) – WONDER GRASS OR WEED?

Cenchrus ciliaris L. (buffel grass) is a deep rooted, summer growing, perennial tussock

grass. A native of southern Africa, India and Indonesia, it was accidentally introduced into

Australia in the 1870s (Paull and Lee 1978). C. ciliaris is well adapted to a wide range of

soils and the climate of arid and semi-arid Australia and recognition of its drought tolerance

and productivity led to its uptake and development by the pastoral industry (Hall 2001). A

number of cultivars have been introduced and their use promoted in arid and semi arid

regions (Humphreys 1967; Cavaye 1991). Cenchrus ciliaris has also been used for erosion

control (Albrecht and Pitts 1997) and rehabilitation of degraded and mined land (Grigg et al.

2000).

Today C. ciliaris is one of the most important pasture introductions in arid and semi-arid

Australia. It is by far the most widely distributed sown pasture grass in Queensland (Cavaye

1991) and has become naturalized over large areas. It is estimated to be abundant or

dominant over 30 to 50 million ha in Queensland (Woinarski 2001) and covers large areas of

Western Australia, Northern Territory, South Australia and New South Wales (Pitts and

Albrecht 2000; Franks and Hannah 2001). In a pastoral context C. ciliaris has been highly

regarded, both for its value as pasture for livestock and for its soil protecting properties (Hall

2001). “Buffel grass has brought great financial benefit to many individual producers and

companies…as well as supported many rural communities, because of its benefit to the

pastoral industries. Its wide adaptation and tolerance of drought, fire and over-grazing have

been major assets” (Hall 2001).

Recently however, the perceived negative effects of this species have gained attention.

Cenchrus ciliaris is reported to be reducing biodiversity and altering fire regimes

(Humphries et al. 1991; Low 1997; Franks 2002; Butler and Fairfax 2003). Its weediness

derives from its dominance of native vegetation and natural dispersal ability (Hall 2001). It is

a strong competitor, producing more biomass than many native perennial grass species, and

its high seed yields and light fluffy seed allow it to spread readily via wind and water. In

particular, C. ciliaris is seen as a major threat to important mesic habitats within the arid

zone (Humphries et al. 1991). These sites are critical parts of the landscape, providing

concentrations of water and nutrient resources and refugia for plants and animals. The

displacement of native vegetation by C. ciliaris in these sites is believed to be threatening the

survival of rare species and altering the food supply of native animals.

1

The spread of C. ciliaris into non-target areas has become a serious concern to non-pastoral

land managers, particularly those responsible for conservation areas. Cenchrus ciliaris is

now considered one of Australia’s worst environmental weeds (State of the Environment

Advisory Council 1996). In contrast to Hall’s (2001) summation, Humphries et al. (1991)

stated: “Buffel grass is one of a large number of introduced grass species which is causing

insidious changes to ecosystem composition, structure and function”. The spread of C.

ciliaris into non-target areas as well as recognition of the need to manage Australia’s

pastoral lands for multiple uses means the status of C. ciliaris – weed or valued pasture plant

– has become highly controversial.

Although the literature relating to biological invasions has grown considerably over the last

decade (Barrow 1995), much is lacking from our understanding of the impacts and

management of invasive species. In the case of C. ciliaris, there is a large literature dealing

with this species in an agricultural context including issues such as variety development,

establishment and animal production. However, there is relatively little published

information relating to the impacts and management of this species as an invasive plant.

Although C. ciliaris has been associated with reduced biodiversity and altered fire regimes,

there are few studies quantifying such impacts. There is even less published information

regarding management strategies for the control of C. ciliaris. Most of the work on these

issues has focused on C. ciliaris in central Australia where this species is perceived to be a

major problem (Griffin 1993; Pitts and Albrecht 2000). The focus of this project is the

impact and management of C. ciliaris as an invasive species in northern Queensland.

Although there is increasing concern regarding C. ciliaris as a threat to semi-arid habitats in

north Queensland, there is relatively little published information regarding its weed status in

these habitats.

1.2. THE ISSUES: IMPACTS AND MANAGEMENT

Two questions relating to invasive species require consideration. Firstly what are the impacts

of invasive species on ecosystem components and processes and, secondly, how can we

manage them?

Whereas many workers cite the negative effects of invasive species, there are relatively few

studies that quantify the effects of exotic plant species on natural and semi-natural

2

environments in Australia (Humphries et al. 1991; Adair and Groves 1998; Vranjic et al.

2000). Australian and overseas studies report that exotic plants are generally associated with

a decline in species richness or diversity (Adair and Groves 1998). Cenchrus ciliaris has

been associated with reduced herbaceous species richness in Hawaii (Daehler and Carino

1998), Mexico (Saucedo-Monarque et al. 1997) and Australia (Fairfax and Fensham 2000;

Franks 2002). However, few studies provide conclusive evidence of causal relationships

between exotic plants and species richness. The impacts of such species must be determined

before control programs are undertaken (Adair and Groves 1998).

“The formulation of management regimes that reduce or prevent further weed invasion and

consequent loss of biological diversity is viewed as one of the most urgent tasks for all

Australian natural ecosystems” (Adair and Groves 1998). Fire is one of the cheapest tools

available for managing vegetation on a large scale. Fire can influence the tree-grass balance

in favour of the grass layer and there is growing recognition of the value of fire for

manipulating invasive shrub and tree species. Although its application for control of fire-

adapted grasses such as C. ciliaris may initially appear inappropriate, manipulation of the

fire regime may offer opportunities to reduce the competitive advantage of C. ciliaris over

native species (Daehler and Carino 1998). The effects of fire are a consequence of complex

interactions between the characteristics of the fire regime and of the vegetation. Research is

needed to determine how fire influences interactions between native and invasive species.

1.3. THIS PROJECT

The overall aim of this project was to enhance our understanding of C. ciliaris as an

invasive, weedy species in northern Queensland. This project focused on two issues: (1) the

impacts of C. ciliaris on herbaceous species richness and (2) the use of fire in the

management of C. ciliaris. These issues are covered separately in chapters two and three.

Chapter two consists of two main sections. Section 2.1 is a review of the impacts of invasive

plants. The effects of invasive plants in general, and C. ciliaris in particular are described.

The controversy surrounding plant introductions in Australia and methods used to assess

invasive plant impacts are also discussed. Section 2.2 describes two field studies

investigating the relationship between C. ciliaris and species richness in the herbaceous

layer.

3

The aim of the first study was to compare herbaceous species richness in habitats in which C.

ciliaris was either dominant or absent. Herbaceous species composition was surveyed in C.

ciliaris-dominated plots and non-C. ciliaris plots at sites in the Dalrymple Shire, north-

eastern Queensland. In such site-comparison studies it is often difficult to determine the

effects of exotic plants on species richness because various factors affecting species richness,

such as site differences in grazing, are often confounded with the presence of the exotic.

The aim of the second study was to investigate the relationship between C. ciliaris and

herbaceous species richness at a single site in the absence of grazing to help minimize some

of these confounding factors. Rather than comparing species richness between areas with and

without C. ciliaris, the relationship between C. ciliaris and herbaceous species richness was

investigated by comparing areas with varying levels of C. ciliaris biomass. Herbaceous

species richness was surveyed in small plots that varied in composition, ranging from plots

dominated by C. ciliaris to plots dominated by a native perennial grass, Bothriochloa

ewartiana. The specific aims of the study were (1) to compare C. ciliaris and B. ewartiana

for their effects on herbaceous species richness and (2) to investigate the effects of scale on

these relationships.

Chapter three describes a series of studies investigating the effects of season of burning on

two C. ciliaris-dominated communities in northern Queensland. The overall aim of the work

was to determine the effects of season of burning on these communities to assess fire as a

tool for managing C. ciliaris-dominated vegetation. Three burning treatments were imposed

in the two communities: early dry season burn (June), late dry season burn (November) and

no burn (control). The experimental sites and fire treatments are described (section 3.3)

following a review of the use of fire as a tool for manipulating vegetation composition

(section 3.2). Individual studies investigating burning effects on establishment site

availability, propagule supply, seedling recruitment and community composition are

described (sections 3.3-3.10) and the implications of the findings for using fire as a tool to

manage C. ciliaris are discussed (section 3.11).

The final chapter of this thesis (chapter four) gives a brief summary of the findings from

chapters two and three and discusses future research needs.

Several cultivars of C. ciliaris are established in Australia. It is not always possible to

identify individual cultivars in the field and I did not determine the identity of the cultivars

present at sites involved in this work. Authorities to species names are provided in

appendices at the end of the relevant sections.

4

CHAPTER 2. CENCHRUS CILIARIS AS AN INVASIVE SPECIES

2.1. LITERATURE REVIEW: IMPACTS OF INVASIVE PLANTS

2.1.1. INTRODUCTION

Biological invasions are seen as one of the major threats to ecosystem integrity both in

Australia and around the world (Heywood 1989; Vitousek et al. 1997). Invasive, exotic

plants have been associated with changes in ecosystem components and processes, resulting

in changes in community composition and structure. Invasive plants alter communities by

direct competition and displacement of native species (Humphries et al. 1991). They also

alter ecosystem processes such as water and nutrient cycling, geomorphological processes,

micro-climate and disturbance regimes (Macdonald et al. 1989; D’Antonio and Vitousek

1992; Gordon 1998). Where invasions result in ecosystem-level effects, ecosystems can be

irreversibly altered and their capacity to provide goods and services can be reduced (Masters

and Sheley 2001). Such effects can have profound economic and cultural consequences

(Vermeij 1996). Invasive plants can destroy wildlife habitat, diminish forest regeneration and

production, render rangeland unsuitable for grazing, increase soil erosion, degrade streams

and lakes and reduce recreational opportunities such as hunting, fishing, camping and

boating (Franklin et al. 1999).

Exotic plants have been introduced into Australia both accidentally and intentionally.

Although many accidental introductions are generally perceived as weeds, the status of

deliberately introduced plants, such as those introduced for pasture improvement, is more

controversial. Until recently, problems associated with such plants had received little

attention (Low 1997). However, there is now growing concern about their negative effects.

In particular, concern has been raised about the negative effects associated with introduced

grasses (Humphries et al. 1991; Adair and Groves 1998). Invasion by exotic grasses has

been associated with displacement of native species, changes in fire regimes and alteration of

wetland structure (Low 1997).

One of the most controversial of these grasses is Cenchrus ciliaris, a drought and grazing

tolerant perennial tussock grass that has become well established in Australia’s arid and

semi-arid zones. In a pastoral context, C. ciliaris has been highly regarded for its value as

pasture for livestock and for its soil protecting properties (Hall 2001). However, its apparent

negative effects on biodiversity and fire regimes are of considerable concern, particularly for

5

management of conservation reserves. It has been associated with reduced plant (Fairfax and

Fensham 2000; Franks 2002) and animal (QPWS 2001) diversity as well as an increase in

the intensity and incidence of fire (Humphries et al. 1991; Butler and Fairfax 2003). Given

these changes and its vast range, C. ciliaris is believed to be having a major impact on

Australia’s biodiversity (Humphries et al. 1991; Woinarski 2001).

Although there is a large literature discussing the ecological effects of invaders, much of this

is purely anecdotal (Parker et al. 1999). This is very much the case for species such as C.

ciliaris in Australia where there are relatively few published data quantifying the perceived

negative impacts of this species. Understanding and being able to quantify the impacts of

invasive species is necessary for developing appropriate management strategies (Adair and

Groves 1998, Masters and Sheley 2001). Given the controversy associated with deliberately

introduced species, understanding and being able to quantify plant impacts is particularly

important. The status of C. ciliaris as a major environmental weed is highly contentious.

While reported to be one of Australia’s most damaging plant imports (Low 1997), it is

considered by many to be a highly desirable and valuable pasture introduction. A better

understanding of this species in an ecological context will provide a basis for rational debate.

2.1.1.1. This review

In this review the phenomena of biological invasions and the impacts of invasive plants are

introduced. The focus is on the detrimental effects of invasive plants in natural and semi-

natural ecosystems rather than invasive plant effects in urban and cropping situations. The

controversy regarding deliberately introduced plants in Australia is discussed with emphasis

on pasture introductions. The introduction and development of C. ciliaris in Australia is

described followed by a review of the reported impacts of this species. Finally, the

methodologies and difficulties of quantifying invasive plant impacts are discussed.

2.1.1.2. A note on terminology

In the expanding field of invasion ecology there is considerable confusion over terms used to

describe various concepts (Richardson et al. 2000). Terms such as ‘exotic’, ‘invasive’ and

‘weed’ may have broadly understood meanings but detailed definitions may be more

problematic. (e.g. see Kendle and Rose (2000) for a discussion of the difficulties of defining

‘native’ and ‘exotic’). As a rule, ecological terminology is “messy” and discrete ecological

6

categories rarely exist (Schwart 1997). However, problems with terminology are perhaps

more severe in invasion ecology than in other disciplines since the notion of invasion

frequently involves anthropocentric concepts such as aggression, intrusion etc. (Richardson

et al. 2000). The usage of some invasion ecology terms is briefly described below.

What are ‘invasive’ plants? Invasive plants have been defined as naturalised plants that

produce offspring, often in very large numbers, at considerable distances from parent plants

and thus have the potential to spread over a considerable area (Richardson et al. 2000). What

then are ‘naturalised’ plants? The term ‘naturalised’ has been used loosely with various

meanings being implied. Although ‘naturalisation’ has been used as a synonym for

‘invasion’, Richardson et al. (2000) argue that these terms represent two overlapping but not

identical phases in the naturalisation/invasion process. They recommend that the term

‘naturalised’ refer to alien plants that reproduce consistently, sustaining populations over

many life cycles without direct intervention by humans, and that they generally recruit

offspring close to adult plants and do not necessarily invade other ecosystems. The possible

differences in interpretation of these definitions demonstrate the difficulties of providing

precise terminology.

Many definitions of the term ‘invasion’ assume or imply negative effects. For example that

of Mack (1996): “Plant invasions describe the proliferation and persistence of a species in a

new range such that it has detrimental consequences (abiotic, biotic, or both)”. However, it

has been suggested that the term be used without reference to any effects (Richardson et al.

2000). That is, invasion occurs “when an introduced species reproduces and expands its

range beyond the site of introduction” (Daehler and Strong 1994). Much of the literature

dealing with the effects of plant invasions focuses on detrimental effects. This is also the

focus of this review. However, it is important to note that, although any successful invasion

must have some consequences for the other species present, most effects are minor

(Williamson 1996). In some instances the effects of invasive plants may actually be desirable

or at least tolerable. For example, invader species may increase surface cover, thereby

reducing soil erosion (Walker and Smith 1997). They may provide feed for livestock or

nectar for honey production during periods when other species are not flowering (Sindel

2000). They may help maintain higher levels of pollinators and act as nursery plants for

native species (Woods 1997). Some aquatic weeds, for example Salvinia molesta, are able to

remove excess nutrients and other pollutants from effluent (Sindel 2000).

The term ‘environmental weed’ needs clarification since it is encountered frequently in

invasive plant literature and various definitions have been proposed (refer Adair (1995) and

7

Csurhes and Edwards (1998)). Whereas some restrict this term to refer to weeds of areas

managed for conservation (e.g. Swarbrick and Skarratt 1994), I accept a broader definition in

which the label ‘environmental weeds’ refers to species that invade native and semi-native

systems including areas set aside for conservation as well as areas managed for other uses. In

this context, C. ciliaris is considered an environmental weed even when it has invaded areas

that are not designated conservation reserves.

2.1.2. BIOLOGICAL INVASIONS

“For as long as humans have travelled over and between land masses, species have

been transported, deliberately or inadvertently, from their native ranges to new,

previously unoccupied areas” (Manchester and Bullock 2000).

Biological invasion happens when an organism reproduces and expands beyond its previous

range (Williamson 1996). Although biological invasions are generally associated with

human activities, they are natural phenomena that long antedate humankind (Daehler and

Strong 1994). All organisms have dispersal mechanisms that provide the potential to

colonize new areas. However, today invasions represent a “new ball game” (Wagner 1993).

Modern human activities have caused the breakdown of barriers to dispersal. The movement

of species is now occurring at rates without precedent in the last tens of millions of years

such that “taxa that evolved in isolation from each other are being forced into contact in an

instant of evolutionary time” (D’Antonio and Vitousek 1992).

The invasion process involves a series of stages that have been variously defined by different

authors. Vermeij (1996) described three successive stages: (1) arrival, or the dispersal of

individuals to the recipient region, (2) establishment, the stage in which the new population

sustains itself through local reproduction and recruitment, and (3) integration, when local

species and the invader respond to each other ecologically and evolutionarily. That is, the

invader modifies species in the recipient community and local species modify the invader.

Groves (1986a) also divided the invasion process into three stages: (1) introduction, (2)

colonisation and (3) naturalization, while Humphries et al. (1991) described five stages (1)

introduction, (2) establishment, (3) survival, (4) production of numerous propagules and (5)

widespread dispersal.

8

Importantly, most arrivals and colonizations do not result in biological invasions (Daehler

and Strong 1994) and most invasions have little impact on the systems they enter. This

generalization has been described by the ‘tens rule’ (Williamson 1996): 10% of introductions

will become established or naturalised and 10% of these will become pests. Williamson

(1996) noted, while that this rule is very rough and ‘pests’ are defined by human perception

rather than by ecological effects, it seems to be true. Consequently, while all invasions alter

the invaded ecosystem, most effects are minor and the invader may simply increase species

richness (Williamson 1996).

However, the dramatic increase in the numbers of species being moved from place to place

significantly increases the numbers of species that may have major impacts. Significant

changes may result where ecosystem processes are sufficiently disrupted by the invasion

(Masters and Sheley 2001). Invasive plants may disrupt ecosystem processes such as

erosion, stream sedimentation, energy flow, nutrient and water cycling, plant regeneration

patterns and fire regimes (Macdonald et al. 1989; D’Antonio and Vitousek 1992; Masters

and Sheley 2001). These impacts can result in displacement of native species, causing

significant changes in community composition and structure. Although biological invasions

may be perceived as less important than other major human impacts such as climate change

and deforestation, D’Antonio and Vitousek (1992) pointed out that biological invasions have

caused more species extinctions than human-caused climate and atmospheric change (more

so for animal invaders, see Macdonald et al. 1989). They also stated that, whereas changes in

climate, atmosphere and land use may be reversible in hundreds to thousands of years, many

of the changes associated with biological invasions must be considered irreversible.

The impacts of invasive plants vary spatially and temporally, depending on the attributes of

the invasive species and the characteristics of the habitat being invaded (Csurhes and

Edwards 1998). Some species, such as the aquatic fern Salvinia molesta, grow aggressively

and disperse extensively so their impacts are relatively immediate and widespread. Other

species, such as Cyperus papyrus, become naturalised at few sites and spread very little so

that their effects are very localised (Humphries et al. 1991). Although many species are

common and have little measurable effect on other species, it is important to note that it may

take a long time for a species to change from innocuous to pest (Williamson 1996). A

spectacular example is Mimosa pigra. This leguminous shrub was introduced into Darwin in

the late 1800s but did not ‘explode’ until the late 1970s (Braithwaite et al. 1989). Williamson

(1996) describes case studies that demonstrate how the impacts of invasive species may

change over time.

9

Invasive plants come from different plant families and vary in life form (Williams and West

2000). However, certain families, for example the Poaceae and Asteraceae, and genera, for

example Bromus, Poa, Cirsium and Centaurea contain a large proportion of the world’s

problem species (Mack 1996). Exotic, invasive grasses are recognized as a problem around

the world, displacing native species, altering water and nutrient cycles, geomorphology and

microclimate, and causing changes to fire regimes (D’Antonio and Vitousek 1992). In

Australia there is increasing concern about the effects of deliberately introduced pasture

grasses (Humphries et al. 1991; Low 1997; Adair and Groves 1998).

Serious attention was first focused on the nature and significance of biological invasions

with the publication of “The ecology of invasions by animals and plants” in 1958 by C.S.

Elton (Barrow 1995). Today biological invasions are recognized as one of the major threats

to nearly all biogeographical regions on earth (Adair and Groves 1998), although the impacts

of biological invasions are perceived to be particularly serious in North America, southern

Africa, Australasia and the oceanic islands such as the Galapagos and Hawaii (Manchester

and Bullock 2000). Consequently, there is a significant research effort directed at issues

relating to biological invasions. These include identifying the characteristics of successful

invaders and the characteristics of communities that influence their ability to be invaded, as

well as determining the impacts of invasive species and developing strategies for preventing

and controlling invasive species.

2.1.3. IMPACTS OF INVASIVE PLANTS

The considerable literature on the effects of invasive species includes studies detailing

specific effects of particular species as well as reviews discussing invasive plant impacts in

broader contexts. However, there appears to be no clear cut way to categorize invasive plant

impacts. Impacts on primary productivity, water and nutrient cycling and species interactions

are linked such that changes in one are likely to cause changes in others (Walker and Smith

1997). These interactions between impacts make them difficult to categorize. As noted by

Williamson (1996), “ecological effects can seldom be categorized neatly into distinct

classes”. For example, D’Antonio and Vitousek (1992) differentiated between ‘competitive

effects’, where the invasive plant utilized resources reducing their availability for other

species, and ‘ecosystem effects’, where the invasive plant altered resource availability at a

‘whole-system’ level by altering ecosystem processes such as nutrient cycling. However,

10

they noted that these categories were not exclusive; competitive interactions can have

ecosystem level consequences and visa versa. For example, grasses may competitively

exclude trees and shrubs and the resultant changes in fuel loads may alter fire regimes,

affecting nutrient cycling. These ecosystem level impacts may alter competitive interactions

by changing resource availability.

Different schemes have been used to categorize invasive plant impacts. For example, Walker

and Smith (1997) discussed the impacts of invasive plants in terms of effects on primary

productivity (vegetation structure, composition, growth and diversity), nutrient dynamics,

soil moisture and salinity, disturbance regimes (fire, erosion, plant-herbivore interactions)

and community dynamics (competition, stability, succession etc.). Parker et al. (1999)

considered impacts at five levels: (1) effects on individuals (e.g. growth, mortality), (2)

genetic effects, (3) population dynamic effects (abundance, growth etc.), (4) community

effects (species richness, trophic structure) and (5) effects on ecosystem processes (nutrient

availability primary productivity etc.). Macdonald et al. (1989) described the impacts of

invasive plants in terms of effects on “ecosystem structure” (i.e. species composition, genetic

diversity) and “ecosystem function” (e.g. nutrient cycling, hydrology, soil erosion,

decomposition). In general, the specific classes proposed by one author(s) do not fit well

with those used by others (Williamson 1996). For example, while D’Antonio and Vitousek

(1992) distinguished between competitive and ecosystem level effects, others, for example

Csurhes and Edwards (1998), included competition as an ecosystem level function.

A recurrent, broad scheme used to categorize invasive plant impacts is that which divides

impacts into those affecting ecosystem structure, that is impacts on ecosystem components

such as individual species and/or communities, and those affecting ecosystem function

including processes such as nutrient and water cycles and disturbance regimes (as in

Macdonald et al. 1989). However, as described above, even these two broad categories

overlap and division between them is not consistent in the literature. I have used this scheme

to summarize the impacts of invasive plants below.

2.1.3.1. Impacts on ecosystem structure/components

Invasive plant impacts on ecosystem structure, or ecosystem components, involve changes to

biodiversity. Biodiversity can be described at three levels (genetic, species and ecosystem

diversity) and invasive plants are capable of affecting biodiversity at one or more of these

levels (Adair and Groves 1998).

11

Genetic diversity

Invasive plants may alter genetic diversity, which is the diversity within a species (Lawrence

1995), via local extinction or hybridisation effects. Invasive plants have been associated with

losses of local populations and it has been suggested that such losses result in reductions in

the genetic diversity of the taxa involved (Adair and Groves 1998). Exotic species may also

result in ‘genetic pollution’ (Robin and Carr 1986) by hybridizing with indigenous taxa. The

resultant hybrids may then out-compete indigenous populations and possibly cause their

elimination (Carr 1988). Robin and Carr (1986) give examples of hybridising species in

Australia, noting that some Grevillea and Acacia hybrids have become environmental weeds.

By hybridizing with wild populations, invasive species introduce greater genetic variability,

increasing vigour and ecological amplitude (Low 1997). Macdonald et al. (1989) provide

examples of invasive plant hybidisation in other parts of the world.

Species diversity

The majority of studies investigating the impacts of invasive plants have focused on the

effects on species diversity (Adair and Groves 1998). Species diversity is the combination of

the number of species present and the way in which the individuals are distributed amongst

the various species (Adair and Groves 1998). Although it is a relatively well-defined concept

and is relatively easily measured, there is some debate about the usefulness of species

diversity indices (Adair and Groves 1998). Many studies actually quantify invasive plant

effects on species richness (the number of species in a given area) rather than on species

diversity. Some studies simply document changes in abundance of specific species rather

attempt to measure impacts on species diversity.

Few invasions cause extinctions and most invasions have little impact on species richness,

apart from simply increasing local diversity (Williamson 1996). However, some invasive

plants have profound impacts on diversity. Invasive plants have generally been associated

with declines in species richness or diversity (Adair and Groves 1998) and negative

correlations have been found between exotic species abundance and numbers of native

species (Bridgewater and Blackshell 1981; McIntrye et al. 1988; Vlok 1988; Batianoff and

Franks 1998). Plant species richness has been reduced in areas invaded by various life forms

including: trees and shrubs (e.g. Cytisus scoparius, Smith 1994), vines (e.g. Clematis vitalba,

Ogle et al. 2000), forbs (e.g. Euphorbia esula, Belcher and Wilson 1989), grasses (e.g.

Agropyron cristatum and Agropyron desertorum, Lesica and DeLuca 1996) and aquatic

12

plants (e.g. Myriophyllum spicatum, Boylen et al. 1999). Nineteen of 20 studies on

environmental weed impact in Australia reviewed by Adair and Groves (1998) reported a

decline in species richness, canopy cover or frequency of native species. The reverse has

been found, where invasive plants are associated with increased species richness. This

generally occurs on disturbed land where invasive plants have acted as ‘nurse plants’ for the

regeneration of other species (Adair and Groves 1998).

The majority of studies investigating the impacts of invasive plants on biodiversity have

focused on changes in floristic composition and less is known about invasive plant impacts

on animal diversity (Adair and Groves 1998). Weed invasions may have a range of effects

on the survival of native animals by altering food supply, nesting sites, cover and protection

from predators (Loyn and French 1991) and both positive (Hedge and Kriwoken 2000) and

negative (Usher 1986) effects of plant invasions on animal communities have been reported

(Adair and Groves 1998). Some studies have found that, whereas species richness appears

unchanged by plant invasion, changes in the abundance of particular taxa are evident. For

example, the abundance of litter invertebrate taxa was found to differ between coastal heath

with and without Chrysanthemoides monilifera (bitou bush) in New South Wales, although

the numbers of invertebrate species were similar in these two habitats (French and Eardley

1997). The invasion of riverine woodland in central Australia by Tamarix aphylla was

associated with changes in the trophic structure of bird assemblages and an overall decline in

bird abundance but no change in bird species richness (Griffin et al. 1989).

In their review of environmental weed impacts, Adair and Groves (1998) concluded that

most studies report negative effects of weed invasion on some measure of biodiversity.

However, it is important to note that the effects of invasion can be positive, negative or

neutral depending on the biotic group measured (Groves and Willis 1999). For example, in

Californian shrublands the invasive perennial tussock grass Cortaderia jubata has been

associated with reduced richness of native shrub species but increased numbers of exotic

plant species such that overall plant species richness is similar to that of non-invaded areas

(Lambrinos 2000). The effects of invasive species also depend on the invaded community.

Belnap and Phillips (2001) found a reversal of the effects of Bromus tectorum invasion on

soil biota between two grassland communities in Utah. Consequently, generalizations such

as “plant invasions…..are typically characterized by a decline in species diversity at all

trophic levels within an ecosystem” (Beerling 1995), are perhaps premature when so little is

known about the impacts of invasive plants on many groups of organisms. It is also

important to recognize that invasive plant impacts may change over time. For example,

although invasion of temperate woodland in New South Wales by Cytisus scoparius (broom)

13

is associated with reduced herbaceous species richness, in the longer term species richness

may recover as the broom canopy begins to open (Smith 1994).

Although many studies document negative impacts of invasive plants on species diversity

and on particular species, there is little information regarding the role of invasive plants in

species extinction. (In contrast, there are many examples of animal invasions resulting in

extinction particularly on islands, see Macdonald et al. 1989). Determining the role of

invasive species in the extinction of species is difficult since (1) it is difficult to ‘prove’ that

a species has, in fact, become extinct and (2) few species are studied in sufficient detail

before they become extinct so the factors responsible for their demise are unknown

(Macdonald et al. 1989). The evidence available suggests that, while invaders (referring here

to both animal and plant invaders) often cause extinction on oceanic islands and in lakes,

they rarely cause extinctions in the sea or on large land masses. Rather, they probably restrict

the ecological range of native species (Vermeij 1996). Macdonald et al. (1989) noted that

there was only one well documented case where invasions by exotic plants posed a major

threat to native plant species diversity at a continental scale: the fynbos and Karoo biomes in

South Africa. There have been no documented cases of continental species extinctions

attributed solely to weed invasion in Australia (Adair 1995). Leigh and Briggs (1992) list

weeds as a major cause in the extinction of only four plant species. However, invasive plants

are currently threatening several species (Williams and West 2000). For example two

endangered species, Pterostylis arenicola and Pimelea spicata, are threatened by the

introduced Asparagus asparagoides (bridal creeper) (Groves and Willis 1999). The

fragmentation and disintegration of native vegetation caused by weed invasion has lead to

range contractions, reduced abundance and a decline in the diversity of native biota. In many

areas, species extinctions have occurred at local and regional levels (Adair 1995). Rare

plants are often threatened by extinction (Daelher and Strong 1994) and numbers of rare

species have been found to be negatively correlated with the abundance of exotic species

(McIntrye and Lavorel 1994).

Ecosystem diversity

Ecosystem diversity is the number of different ecosystems in a given environment

(Lawrence 1995). The impacts of plant invasion on ecosystem diversity are generally not

explicitly quantified since determining whether invasion-induced changes represent

increases or decreases in ecosystem diversity depend on the scale of concern and how the

ecosystems are defined. Rather, the threats of invasive species to particular ecosystems are

recognized at various scales.

14

Where invasive plants threaten ecosystems, it is usually where the invading species creates a

new stratum and/or where essential ecosystem functions are altered (Adair and Groves

1998). Invasive plants may have dramatic effects on vegetation structure. For example

Mimosa pigra converts sedgeland to tall shrubland. Acacia nilotica converts dry grassland

into tall shrubland (Panetta and Lane 1996) and is considered a threat to Mitchell grasslands

Australia wide (Humphries et al. 1991). Vine species such as Thunbergia grandiflora,

Macfadyena unguis-cati and Anredera cordifolia convert forest vegetation into vine thicket

(Panetta and Lane 1996). Cryptostegia grandiflora (rubber vine) threatens native vine

thickets, gallery forests and dry rainforest in the monsoonal belt of northern Australia

(Humphries et al. 1991). Humphries et al. (1991) list a number of Australian ecosystems at

risk from plant invasion. Invasive plants also affect the diversity of ecosystems elsewhere.

For example, replacement of native flora in California has been so thorough that a new

vegetation type ‘valley grassland’ has been created (Heady 1977).

2.1.3.2. Impacts on ecosystem function/processes

Invasive plants may affect ecosystem processes. These have been defined as “whole-system

fluxes of energy, the amounts and pathway of inputs, outputs and cycling of materials and

the ways that these vary in time” (D’Antonio and Vitousek 1992). They include water and

nutrient cycling, productivity, geomorphological processes, and disturbance regimes

(Williamson 1996).

Vitousek (1990) suggested that invasive species could have ecosystem-level effects where

they (1) differ substantially from natives in resource acquisition or utilization; (2) alter the

trophic structure of the invaded area; or (3) alter disturbance frequency and/or intensity.

Invasive plants may alter ecosystem properties where they add a life form that is not well-

represented in the native flora (Vitousek 1986) and/or where they add a new biological

process, such as nitrogen fixation, to a system (Vitousek et al. 1987).

Whereas numerous studies demonstrate that biological invasions can alter population

dynamics and community structure, there is much less information available regarding

ecosystem-level effects (Ramakrishnan and Vitousek 1989). Vitousek (1990) considered that

relatively few invasive species have ecosystem-level effects. However, Adair and Groves

(1998) concluded that the numbers of species altering ecosystem functions were likely to be

higher than that reported because this issue is not widely researched. Gordon (1998) also

15

suggested that ecosystem-level effects were more common than thought, but noted that much

of the current information was anecdotal. It is difficult to demonstrate ecosystem-level

effects of invasive species because the impacts of invasion are often confounded with

ecosystem disturbance (Ramakrishnan 1991), making it difficult to separate out the effects of

exotic species from the effects of disturbance that allows the species to establish (Vitousek

1986). The impacts of environmental weeds on ecosystem processes have not been well

researched in Australia (Williams and West 2000) and there are few quantitative overseas

studies on ecosystem-level impacts of invasive plants (Adair and Groves 1998).

Where invasions alter ecosystem properties, there may be major changes in ecosystem

structure and function. Ecosystem-level changes in fluxes of water or energy and/or the

cycling and loss of material can alter conditions for life for all the organisms in an ecosystem

(Ramakrishnan and Vitousek 1989). As pointed out by Vitousek (1990), exotic species that

alter ecosystem properties do not merely compete with or consume native species; they alter

the fundamental rules for existence for all organisms in the area. Consequently, ecosystem

effects may be more serious than local extinction since they lead to the severe disruption of

many species rather than the local loss of a few (Williamson 1996). In fact, some invasive

plant species have been described “transformer” species since they become dominant and

change the character and condition of ecosystems over substantial areas (Wells et al. 1986).

The alteration of some ecosystem processes can have repercussions for others, thereby

increasing the scale of change (Vitousek 1986). Positive ecosystem-level feedback may

occur where there are changes in nutrient availability or disturbance regimes (Vitousek

1986). For example, invasion by weedy plants in Hawaii favours feral pig activity, which in

turn favours the spread of the weeds via the effects of soil disturbance and seed spread

(Smith 1985). Species whose invasive capacity involves such positive feedback present the

greatest threats to ecological systems (Levine et al. 2003). The effects of invasive plants on

hydrological cycles, geochemical processes, geomorphological processes, primary

productivity and disturbance regimes and community dynamics are briefly described below.

Hydrological cycles

Some invasive species alter water distribution. For example, Orbea variegata, an exotic

plant invading chenopod shrublands in South Australia, has a dense root system that is

thought to trap water in the surface soil, reducing movement to lower layers (Dunbar and

Facelli 1999). Changes in transpiration rates associated with invasive species may also affect

hydrology. Sites dominated by the exotic grass Andropogon virginicus, which has invaded

16

tropical rainforest in Hawaii, become water logged as this grass has much lower transpiration

rates than the original native forest vegetation (Mueller-Dombois 1973). Some broadleaf,

invasive plants have lower leaf areas than the annual and perennial grasses they are replacing

and this can alter water capture at invaded sites. Surface water runoff and stream sediment

yields were 56 and 196% higher respectively in a Centaurea maculosa-dominated site

compared with adjacent native perennial grassland (Lacey et al. 1989). In contrast, invasive

species with much higher water usages than the native species they have replaced can reduce

water tables and dry up water holes. For example, Tamarix spp., which have invaded water

courses in North America (Vitousek 1986) and Australia (Griffin et al. 1989), have deep

penetrating roots and can tap into the water table throughout much of the year. The resulting

high transpiration rates can lower water tables (Loope et al. 1988).

Geochemical processes (nutrient cycling)

Invasion by woody and herbaceous species has been associated with altered nutrient cycling

and distribution patterns. Invasion of grasslands by woody plants may create patches of

higher fertility. The deeper root systems of trees and shrubs, together with their relative

longevity, enable them to extract nutrients from soil layers beyond the rooting depth of

herbaceous species. Nutrient and organic matter levels beneath their canopies are increased

via litter drop, canopy drip and stem flow (Johnson 1986). Alternatively, it has been

suggested that the deep root systems of some invasive plants mean that they contribute less

organic matter near the soil surface (Olson 1999). Invasion by nitrogen fixing plants in

particular may result in altered nutrient cycling and soil nitrogen status (e.g. Myrica faya,

Vitousek and Walker 1989).

Effects on nutrient cycles have been attributed to changes in litter quality and quantity

associated with invasive plants. Invasion of exotic Acacia spp. in the South African fynbos

increases soil phosphorus levels since the Acacia spp. have higher phosphorus levels in their

litter, greater litter accumulation and more rapid litter turn over rates than the native species

(Witkowski and Mitchell 1987). The invasive grass Bromus tectorum has been associated

with reduced inorganic nitrogen availability. This grass produces more litter than native

species and its significantly higher C:N and lignin:N ratios decreases N mineralization rates

(Evans et al. 2001).

Invasive plants may affect soil biota as well as soil physical and chemical properties and

these changes interact to affect nutrient cycling and the soil environment. For example

17

Mesembryanthemum crystallinum (iceplant), an invasive species in South Australia and

California, accumulates salt in its tissues and leachate from its residues raises the salinity of

the top soil, creating an unfavourable environment for the establishment or growth of other

species (Kloot 1983). Allelopathic effects have been attributed to some invasive plants (e.g.

Pittosporum undulatum (Gleadow and Ashton 1981) and Lantana camara (Achhireddy and

Singh (1984)).

Geomorphological processes

Some species have landscape transforming abilities via their effects on geomorphological

processes. Both woody and herbaceous invasive species can have major impacts on the

physical environment by altering sediment deposition patterns. Aquatic habitats may be

transformed by species such as Hymenachne amplexicaulis (Houston and Duivenvoorden

2002) that tend to form dense, monospecific stands (Macdonald and Frame 1988). Invasive

Tamarix spp. may alter river courses as they form dense stands that increase sedimentation

rates by trapping and stabilizing sediments (Griffin et al. 1989). Spartina spp. (cord grasses)

can rapidly alter the character of estuaries by increasing sediment deposition. Their growth

along river banks and tidal channels can restrict water flow and cause widening of the flood

plain (Asher 1991).

Invasive species also alter the geomorphological dynamics of beaches and dunes. For

example, invasion by the European beach grass Ammophila arenaria has been found to alter

dune formation processes in North America (Mooney et al. 1986) and Australia (Heyligers

1985). The Australasian tree Casuarina equisetifolia has also been found to affect dune

dynamics in subtropical coastal regions of North America (Barbour and Johnson 1977).

Primary productivity and disturbance regimes

Invasive species may increase, decrease or have neutral effects on primary productivity

(Walker and Smith 1997). Plant invasion may increase productivity by providing new life

forms, new phenological patterns and/or new modes of resource acquisition. For example,

productivity has increased dramatically in young volcanic sites in Hawaii that have been

invaded by the nitrogen fixing tree Myrica faya (Walker and Smith 1997). Whereas the

vigorous growth of many invasive species increases productivity (e.g. Spartina spp., Daehler

and Strong 1996), slower growth or growth at a similar rate to the native species may lead to

18

negative or neutral effects on productivity (Walker and Smith 1997).

Invasive species that alter disturbance regimes can have significant ecosystem-level

consequences (Vitousek 1990) and probably the most dramatic impact of invasive species is

their alteration of fire regimes (Woods 1997). Invasive species can alter the rate of spread,

the probability of occurrence, and the intensity of fire, and where they do, they generally

increase fire frequency and/or intensity (D’Antonio 2000).

Invasive species affect fire regimes by altering primary productivity, and thus fuel loads.

Invasive species often produce more biomass than native species, providing higher fuel loads

(Woods 1997). Fire regimes may also be altered by differences in the flammability,

phenology or structure of the invaded community compared with the native vegetation

(D’Antonio 2000). For example, differences in phenology between native and exotic species

increase the fire risk in desert shrublands in Nevada where, in contrast to the native

vegetation, the invasive grass Bromus rubens is able to germinate in spring and produce

biomass that provides fuel throughout summer (Beatley 1966). The invasion of fire-tolerant,

exotic plants often results in positive feedback loops between fire and weed invasion: as the

weed becomes more abundant, fire frequency and intensity increases, fire-sensitive native

species decline and the exotic species increases in abundance, further altering the fire regime

(Vitousek et al. 1997; Rossiter et al. 2003). These invasive species are sometimes referred to

as ‘fire weeds’ due their promotion of fire (Wilson and Mudita 2000).

Fire regimes are affected by various plant life forms. However, increased fire frequency and

intensity have been particularly associated with invasion by exotic grasses (D’Antonio and

Vitousek 1992; D’Antonio 2000). Many studies report increased fire frequency and/or

intensity in association with invasion by exotic grasses (refer D’Antonio 2000). For example,

invasion of Artemisia rangelands in the United States by Bromus tectorum (downy brome)

has led to a reduction in fire interval from 60 to 110 years before invasion to less than five

years after invasion (Whisenant 1990a). Exotic grasses that behave this way in Australia

include Andropogon gayanus, C. ciliaris, Melinis minutiflora and Pennisetum polystachion

(Low 1997).

Other plant growth forms have also been associated with altered fire regimes. In Africa, the

vine Chromolaena odoratum, which rapidly smothers all types of native vegetation and is

prominent in ecotones between forest and grassland, burns readily even when green and can

allow fire to penetrate forest canopies that would otherwise not burn (Macdonald 1983).

Invasion of dry rainforest remnants in Queensland by Lantana camara alters fuel

19

characteristics and increases the flammability of these systems (Fensham 1996). In contrast,

van Wilgen and Richardson (1985) concluded that invasive shrubs decrease the frequency

and intensity of fire in South African fynbos under normal conditions. D’Antonio (2000)

listed some tree and shrub species also considered to have the potential to decrease fire

frequency.

Altered fire regimes may lead to significant changes in community composition as some

species are favoured by the modified fire regime while others may suffer recruitment failure

or are killed (Csurhes and Edwards 1998). Changes in fire regime can result in major

structural changes such as the conversion of forest to savanna or grassland or of

Mediterranean-climate shrubland to grassland (refer Woods 1997). Invasive plant-fire

interactions have been reviewed by D’Antonio and Vitousek (1992) and D’Antonio (2000).

Plant invasions may also affect other disturbance processes including the frequency and

intensity of wind throw, erosion and herbivory (Woods 1997). Invasive vines can change the

likelihood of windfall by binding trees together and weighting their canopies (Thomas 1980).

The establishment of Larrea tridentata (creosote bush) in desert communities in New

Mexico reduces perennial herbaceous cover and soil aggregate stability, leading to soil loss

by wind and water (Whitford et al. 2001). The invasive grass Agropyron cristatum also

increases the erosion risk as it is associated with fewer water-stable aggregates and greater

amounts of exposed soil than native prairie (refer Lesica and DeLuca 1996). Erosion may be

promoted where invasive trees are shallow rooted or more susceptible to burning than the

species they replace (Versfeld and van Wilgen 1986). Invasive plants may also alter plant-

herbivore interactions through changes in plant density, chemistry and seed production

(Walker and Smith 1997).

Community dynamics (competitive interactions, recruitment processes and succession)

Invasive plants may alter species interactions including competition and facilitation as well

as community stability and successional pathways (Walker and Smith 1997).

Woods (1997) identified two main types of competition associated with invasive plants.

Invasion could initiate competition between species of similar life history, resulting in the

reduction or displacement of one or a few directly competing species. Alternatively, the

invading species may have broad competitive effects on species of different guilds. In this

case, the competitive effects may lead beyond floristic replacement to alteration of

20

community structure. Invasive plants are often superior competitors for soil nutrients and

water (D’Antonio and Vitousek 1992). Shading by invasive species, which inhibits growth

and prevents regeneration, is also significant (Leigh et al. 1984) and many studies suggest

that shading is the predominant mechanism responsible for invasive plant effects

(Braithwaite et al. 1989; Downey and Smith 2000). Recruitment patterns may be affected by

altered litter quantity and quality. Greater litter production by invasive species has been

associated with altered recruitment patterns as germinants are smothered or the flammability

characteristics of a site are altered (Vranjic et al. 2000).

The vigorous growth of invasive species such as Spartina spp. (Daehler and Strong 1996)

and Clematis vitalba (Ogle et al. 2000) can smother native vegetation and reduce or prevent

the establishment of native species. However, invasive species may affect recruitment and

successional processes via mechanisms other than direct competition. For example, M. faya

forms dense forests that may alter wind dispersal of propagules (Walker and Smith 1997).

Some invasive species act as hosts for insect and fungal pests that attack native plants (Leigh

et al. 1984) while others alter successional processes via allelopathic effects (Rice 1972).

Shifts in successional pathways may be difficult to detect since they generally occur over

considerable periods of time (Adair and Groves 1998).

2.1.3.3. Mechanisms underlying invasive plant impacts

Processes such as competition, allelopathy and production of flammable biomass have been

defined as ‘mechanisms’ that generate invader impacts such as reduced diversity and

increased fire frequency (Levine et al. (2003), although, as noted earlier, the differentiation

between invasive plant effects and mechanisms producing effects is not clear cut. In general,

the role of these processes in producing invasive plant impacts is poorly understood (Levine

et al. 2003).

Relatively few studies investigating invasive plant impacts on plant community structure

attempt to determine the mechanisms by which such effects are occurring (Levine et al.

2003). However, they often assume competitive effects. “Commonly introduced species are

inferred to be superior competitors to natives for resources” and many studies simply “report

a correlation between increased prominence of the alien and decreased native species

population size, diversity, or vigour” (Daehler and Strong 1994). Competitive effects are

often difficult to verify (Woods 1997) and are usually not investigated. Studies investigating

invasive plant effects on higher trophic levels are generally also comparative (that is

21

comparing invaded versus non-invaded areas), but are more likely to suggest a mechanism

responsible for the observed effect (Levine et al. 2003). For example, Griffin et al. (1989)

suggested that the reduced numbers of birds and reptiles associated with Tamarix aphylla

invasion of river habitat in central Australia was due to lack of food supply and nesting

holes, which made conditions less suitable for birds, and reduced numbers of logs and litter

and a change in litter type, which made tamarisk-invaded areas less suitable for reptiles.

Whereas studies on the impacts of invasive plants on community structure rarely investigate

the mechanisms of impact, studies examining invasive plant effects on ecosystem processes

are generally more mechanistic, often attributing impacts to differences in functional traits

between the invader and the resident species (Levine et al. 2003). For example, the

importance of photosynthetic rates, water relations and growth of invasive African grasses in

affecting ecosystem processes has been discussed by Williams and Baruch (2000). Other

examples are given in Levine et al. (2003).

2.1.4. GRASSES: A FAMILY OF PROBLEM PLANTS

Grasses are a particularly problematic group of invading species. From work by Holm et al.

(1977), Heywood (1989) identified the grass family (Poaceae) as one of three families (the

others being Asteraceae and Papilionaceae) containing notable concentrations of invasive

species of agriculture. The Poaceae is also over-represented among natural area invaders

(Daehler 1998). Of the 18 species listed by Holm et al. (1977) as the world’s worst

agricultural weeds 10 were grasses. Of the 18 species listed as Australia’s worst

environmental weeds, six were grasses (Humphries et al. 1991).

Invasive grasses have been associated with reduced biodiversity (Houston and

Duivenvoorden 2002; D’Antonio and Vitousek 1992; Lesica and DeLuca 1996). They are

good competitors against both herbaceous and woody species (refer D’Antonio and Vitousek

1992) and may displace native species, forming monocultures (Humphries et al. 1991). They

effectively compete for water and nutrients (D’Antonio and Vitousek 1992) and rapidly

growing grasses can reduce light levels at the soil surface (refer D’Antonio and Vitousek

1992) and alter light quality, affecting the growth of other plants (Thompson and Harper

1988). Some invasive grasses have been reported to have allelopathic effects (Rice 1972).

22

As well as directly out-competing native species, invasive grasses affect ecosystem

processes including water (Mueller-Dombois 1973) and nutrient (Ley and D’Antonio 1998)

cycling, geomorpholocial processes such as wetland sedimentation (Daehler and Strong

1996) and dune formation (Heyligers 1985), as well as microclimate and disturbance regimes

(D’Antonio and Vitousek 1992; Williams and Baruch 2000).

By out-competing native species and/or by altering ecosystem processes, invasive grasses

may have ecosystem-level effects resulting in major changes in community composition and

structure. Species such as Hymenachne amplexicaulis (Houston and Duivenvoorden 2002)

and Spartina spp. (Daehler and Strong 1996) out-compete and smother native vegetation,

changing wetland structure. Brachiaria mutica (Para grass) may cause changes in

topography by accumulating large amounts of organic matter (Brown and Ramsay 1999). In

arid and semi-arid habitats grasses are strong competitors against woody species (D’Antonio

et al. 1998). By suppressing woody seedling survival, invasion by exotic grasses may result

in conversion of woodland into grassland.

Probably the most significant impacts of exotic grasses are their effects on fire regimes

(D’Antonio and Vitousek 1992). Invasion by grasses is associated with increased fuel loads

and altered fire regimes in habitats around the world (Csurhes and Edwards 1998). Invasive

grasses have introduced fire into areas where it was previously rare or absent (Smith 1985)

and, in habitats where fire is already a natural occurrence, grass invasion may result in

increased fire frequency and intensity (Levine et al. 2003).

Many exotic grasses have evolved with fire and have mechanisms for surviving and

recovering rapidly after fire (Daubenmire 1968; Vogl 1975). Where exotic grasses are fire-

adapted, they may be promoted by fire at the expense of less fire-adapted native species. For

example, fires promote the noxious, perennial grass Imperata cylindrica by stimulating

rhizome sprouting and by preventing the growth of native species that would otherwise

shade it (Friday et al. 1999).

In addition to being adapted to fire, many invasive grasses actually promote fire and various

attributes explain why this is so. The greater biomass production of invasive grasses

compared with native species is often cited as a causal factor (Williams and Baruch 2000;

Rossiter et al. 2003). Grass fuel distribution and flammability are also important. For

example, the Asian grass Imperata cylindrica, which has invaded sand hills in Florida,

provides a more continuous and higher fuel bed than the native vegetation, resulting in

higher fire intensities (Lippincott 2000). Fire-promoting grasses such as Melinis minutiflora,

23

Andropogon virginicus and Schizachyrium condensatum that have invaded Hawaiian

ecosystems maintain very high dead:live biomass ratios throughout most of the year and

burn at very high moisture contents, thereby providing a continuous fuel supply (Hughes et

al. 1991). Differences in community structure due to grass invasions may also affect fire

regimes. For example, wind speed, which directly affects fire characteristics, may be higher

over grassland than over woodland (Freifelder et al. 1998).

Changes in fire frequency and intensity can have profound effects on community

composition and structure. The seedlings of many woody species are susceptible to fire

(Dyer et al. 1997) and invasive grass-induced changes to fire regimes can reduce or

eliminate woody species. For example, the increased fire frequency associated with Bromus

tectorum invasion in the Great Basin of North America has resulted in changes in vegetation

structure with decline in shrub species (Billings 1990). Fires promoted by invasive grasses in

seasonal sub-montane areas in Hawaii have resulted in reduced cover and diversity of shrubs

and trees (Hughes et al. 1991). Altered fire regimes may have impacts on other ecosystem

components and processes such as nutrient cycling and erosion processes.

Grass species have been moved both actively and accidentally around the world, resulting in

some of the most destructive and widespread invasions (refer D’Antonio et al. 1998). In fact,

grass invasions are considered widespread and effective enough to alter regional and even

global aspects of ecosystem function (D’Antonio and Vitousek 1992). In their review on

invasive grasses, D’Antonio and Vitousek (1992) suggested that at a regional scale invasive

grasses may cause changes to climate, via conversion of forest to grassland, and changes to

atmospheric composition, via changes in fire regime. They suggested that, at a global scale,

grass invasions could contribute to functional change if exotic grass-fuelled fires added

significantly to concentrations of greenhouse or ozone destroying gases, although they noted

that this contribution appeared to be relatively small.

Grass invasions can be found on all continents, although examples from Eurasia and Africa

are relatively rare (D’Antonio and Vitousek 1992). In North America, invasion has been

most severe in the arid and semi-arid west and invasive species include European annual

grasses and perennial bunchgrasses of African, Eurasian and South American origin.

Although introductions of many European species were largely unplanned, perennial species

have been deliberately introduced for pasture and for soil protection (D’Antonio and

Vitousek 1992; Lesica and DeLuca 1996). Tropical Africa has been the centre of origin for a

number of sown forage grasses that, once introduced into America, “have proven to be

explosively aggressive, invading and holding vast areas wherever they have received

24

minimal support by man” (Parsons 1972). Grasses of African origin, also introduced for

pasture, have invaded savannas in Central and South America (D’Antonio and Vitousek

1992) and have become a problem in nature reserves in Brazilian savannas (Pivello et al.

1999). Exotic perennial grasses of African origin are also common throughout Oceania

(D’Antonio and Vitousek 1992) and various species have had major impacts on Hawaiian

native communities (Hughes et al. 1991). In Australia, there are about 310 exotic grass

species (Michael 1994) with European annuals and African bunch grasses predominating

(D’Antonio and Vitousek 1992).

2.1.5. INVASIVE PLANTS IN AUSTRALIA

The history of plant introductions into Australia prior to European settlement is unclear,

although it is believed that at least three plant species (Datura leichhardtii, Solanum

erianthum and Centratherum punctatum) were introduced prior to European settlement

(Jacobs 1981). Undoubtedly most plant introductions have occurred since 1788 and these

have been both accidental and intentional (Leigh et al. 1984). Groves (1986b) and Fox

(1995) have reviewed exotic introductions into Australia.

The rate of plant introductions into Australia is unknown and the numbers of naturalised

plants in Australia are also difficult to determine (Michael 1994). It is estimated that at least

1226 species have become naturalised in Queensland (refer Csurhes and Edwards 1998).

However, despite the fact that only a small percentage of plant introductions become

invasive weeds, exotic plants represent a significant threat to many Australian ecosystems

(Humphries et al. 1991; Adair and Groves 1998). Williams and West (2000) stated that

environmental weeds affect virtually all vegetation communities in Australia.

The majority of Australia’s most serious weeds have been intentionally introduced (Panetta

and Scott 1995). Many new species of grasses, legumes and other fodder species have been

imported for pasture ‘improvement’ with inadequate consideration given to their potential

impact on native ecosystems (Csurhes and Edwards 1998). In his review of grass and legume

pasture introductions in northern Australia, Lonsdale (1994) reported that, of 463

introductions, 60 species (13%) became listed as weeds. Only 21 species (5%) were useful

but some of these were also considered weedy (e.g. C. ciliaris), such that only four species

(less than 1%) were useful and not weedy. The issue of deliberate plant introductions has

25

attracted recent attention as problems associated with such species have become apparent

and it is now recognized that a significant proportion of introduced species has weed

potential.

Introduced grasses in particular have become the focus of attention for government agencies

and others involved with resource management, particularly those managing land for

conservation purposes. In their review of environmental weeds in Australia, Humphries et al.

(1991) considered grasses to be an insidious and serious problem. They noted that grass

invasions are inconspicuous and the mechanism of altering ecosystem structure and function

is gradual so the process is easily overlooked. Adair and Groves (1998) stated that “Exotic

grasses can alter ecosystem functions thereby causing substantial and often irreversible

changes to native biotic communities”.

The impacts of invasive grasses have been described above (section 2.1.4) and have been

reviewed in an Australian context by Humphries et al. (1991) and Low (1997). Low (1997)

considered the two outstanding problems associated with introduced pasture grasses in

Australia to be (1) the impacts of terrestrial species on fire regimes which cause habitat

change and (2) the impacts of semi-aquatic species on wetland structure. Many introduced

grasses are highly productive and produce more biomass than the native species they are

replacing. In addition, they may be highly flammable (e.g. Melinus minutiflora, Paul

Williams, pers comm.) and/or cure later (e.g. Pennisetum polystachyon, Gill et al. 1990),

creating the potential for severe, late season fires. They regenerate quickly after fire and can

form dense monocultures, displacing native species and can become dominant over large

areas (Low 1997). The vigorous growth of semi-aquatic species also displaces native species

and can alter wetland structure (e.g. Brachiaria mutica, Brown and Ramsay 1999 and

Hymenachne amplexicaulis, Houston and Duivenvoorden 2002). Some introduced grasses

considered to be problematic in Australia are listed in Table 2.1.

Attempts to deal with the threat of intentionally introduced species must address conflicts of

interest within society (Panetta and Scott 1995). For example, introductions involving

ornamental or aquarium plants involve conflicts between commercial and environmental

interests. However, the most difficult conflicts to resolve are those involving the introduction

of species for the purpose of increasing primary production (Panetta and Scott 1995).

26

Table 2.1. Some introduced grasses considered to be problematic in Australia.

Achnatherum spp. Ehrharta calycina (Veldt grasses) Andropogon gayanus (Gamba grass) Glyceria maxima Agrostis capillaries (Brown-top vent) Hymenachne amplexicaulis Ammophila arenaria (Marrum grass) Hyparrhena hirta (Tambookie grass) Anthoxanthum odoratum (Sweet vernal) Melinus minutiflora (Molasses grass) Avena spp. Nassella spp./ N. trichotoma (Serrated tussock) Brachiaria decumbens Panicum maximum (Guinea grass) Brachiaria mutica (Para grass) Pennisetum clandestinum (Kikuyu) Briza maxima (Large quaking grass) Pennisetum polystachyon (Mission grass) Bromus diandrus (Great brome) Spartina spp. (Cord grass) Cenchrus cililaris (Buffel grass) Sorghum halepense (Johnson grass) Cortaderia spp. (Pampas grasses) Stenotaphrum secundatum (Buffalo grass) Echinochloa polystachya (Aleman grass) Sporobolus pyramidalis (Giant rats tail grass) Eragrostis curvula Themeda quadrivalvis (Grader grass) (Source: Humphries et al. 1991; Gardener and Sindel 1998; Parsons and Cuthbertson 2001;

CRC for the Sustainable Development of Tropical Savannas web site).

2.1.5.1. Weeds or wonder plants

“Conflicts are likely to arise where plant introductions with high potential value to

rural industries may have an undesirable impact on native ecosystems” (Adair 1995).

Exotic plants have been introduced into many countries to enhance livestock productivity.

Any negative consequences of such introductions were generally unforeseen or were

overshadowed by the perceived benefits of these species. The enthusiasm for such

introductions was significant. One observer commenting on the introduction of Panicum

maximum (guinea grass) to Columbia called P. maximum “a true miracle grass” and

suggested that the unknown person responsible for its introduction into the Magdalena

Valley in the 1830s deserved a statue “as high as New York’s Statue of Liberty, illuminated

by night….so as to be visible throughout the vast area of the new haciendas of the tierra

calienta that it had made productive” (Rivas 1946).

In Australia during the mid 1900s considerable effort was made to source and establish

exotic pasture species (Mott 1986). More recently however, the enthusiasm for introduced

species has been dampened by recognition of their potential negative effects. There is now

considerable debate about the value of pasture introductions (Lonsdale 1994; Low 1997;

McIvor et al. 2000).

27

Many introductions were promoted before their negative effects became apparent. For

example, the invasive tree Tamarix aphylla was “widely and heartily recommended” for

ornamental and shade planting in central Australia (Griffin et al. 1989). However, some

species have been actively promoted by particular agencies despite concerns being raised by

other groups. For example, Low (1997) reported that while there were concerns that

Hymenachne amplexicaulis was “behaving badly” and that, as a high-risk species, it should

be declared a noxious weed, it was being promoted by the Queensland Department of

Primary Industries. Andropogon gayanus was promoted by agencies in northern Australia

who initially ignored warnings regarding its detrimental effects and continued to promote its

use despite considerable opposition (Whitehead and Wilson 2000).

The fact that many pasture introductions have the potential to become weeds should come as

no surprise since some of the characteristics they were selected for, as desirable pasture

species, parallel those suggested as being characteristics of an ideal weed (Mott 1986).

Resilience to disturbance, ability to invade new sites and to out-compete native plants under

grazing conditions are highly valued characteristics, which at the same time put native

systems at risk (Whitehead and Wilson 2000). As summed up by Panetta and Scott (1995):

“Herein lies the crux of the problem: some of the characteristics that make a species useful in

an agronomic context also predispose it to invasive behaviour”.

It is not unusual for an introduction to be welcomed by some interests while condemned by

others (Williamson 1996). For example, in Australia Echium plantagineum takes over

grazing lands and is known as ‘Paterson’s curse’ to graziers. However it is valued by bee

keepers who refer to it as ‘Salvation Jane’ (Williamson 1996). The invasive, stoloniferous

grass Bothriochloa pertusa (Indian couch), which has become naturalised in north

Queensland, is variously regarded as a useful forage grass and a weed of pastoralism (Grice

2000).

Cenchrus ciliaris is a particularly controversial introduced grass. “It seems that some people

love it and others hate is and there are very few in between” (White 1997). It is recognized as

one of the most important pasture introductions for arid and semi-arid Australia and has been

called “the outstanding improved pasture grass in the drier areas of Queensland” (Paull and

Lee 1978). It is very drought-tolerant, can withstand heavy grazing once established, and is

able to establish, persist and produce under harsh climatic conditions (Paull and Lee 1978). It

is a strong competitor, producing more biomass than native perennial grasses and high yields

of light, fluffy seeds that are readily spread by wind and water (Hall 2001). However, these

very characteristics give it its weed potential. It has spread into non-target areas (Griffin

28

1993; Woinarski 2001) and has been reported to form monocultures, displacing native

species and altering fire regimes (Humphries et al. 1991; Franks et al. 2000).

Consequently, different interest groups have very different perceptions of the value of this

species. While ‘Landcare’ supports the pastoral industry to meet its sustainable industry

goals by rehabilitating land with C. ciliaris, a few kilometres away conservation agencies are

concerned with the impact of C. ciliaris in parks (Stafford Smith 1994). The following

transcript from the ABC daily National Rural News program (18 August 1997,

www.abc.net.au/rural/news_states/nrn/nrn-18aug1997-2.htm) illustrates how emotive the C.

ciliaris issue can be.

Cattle producers are outraged at news of a research study underway in Western

Australia that they claim is looking into the possibility of finding a biological control

agent to eradicate buffel grass……. Richard Golden from the Cattlemens Union …

says buffel grass has an estimated worth of $1.5 billion. Richard Golden: We do have

to accept the fact that buffel grass is of vital economic importance to this entire

Australian community, we're not just talking pastoralism here, we're talking about

something which, if it was biologically eradicated, stands a chance of destroying

profitable pastoralism across a whole heap of this country.

2.1.6. CENCHRUS CILIARIS

The genus name Cenchrus is derived from ‘kenchros’ which is the Greek name of a small

millet or one of the cereals used by the ancient Greeks that resembled small millet (Wagner

et al. 1990). There are about 30 Cenchrus species world wide, found mainly in warm, dry

regions of Africa, America and south west Asia (Harden 1993). Eleven species are listed by

Sharp and Simon (2002) as being present in Australia. These are: C. biflorus (=C. barbatus),

C. brownii, C. caliculatus (=C. australis, Hillside burrgrass), C. ciliaris, C. echinatus

(Mossman River grass), C. elymoides, C. incertus (=C. pauciflorus, C. tribuloides, Spiny

burrgrass), C. longispinus (=C. pauciflorus, C. tribuloides, Gentle Annie, Innocent weed), C.

pennisetiformis (Cloncurry buffel, (Paull and Lee 1978)), C. robustus and C. setiger

(Birdwood grass). Only C. caliculatus, C. elymoides and C. robustus are native (Sharp and

Simon 2002).

29

2.1.6.1. Origin, introduction and establishment in Australia

Cenchrus ciliaris is native to north tropical and south Africa, India and Indonesia

(Humphreys 1974). It is believed to have first been introduced into Australia accidentally,

being brought to Wallal in the north-west of Western Australia as a contaminant of Afghan

camel harness between 1870 and 1880 (Marriott 1955). It rapidly became naturalised there

(Humphreys 1974) and its spread to other areas was systematically encouraged (Humphreys

1967). In 1910, seed was spread at Port Hedland and, after World War I, the Western

Australian Department of Agriculture distributed lines of Cenchrus sent by General

Birdwood from Afghanistan (Humphreys 1967). This was the source of the first Cenchrus in

Queensland, arriving in the 1920s (Hall 2001). The early history of C. ciliaris is described by

Humphreys (1967) and its spread in Queensland is described by Cavaye (1991).

Cenchrus ciliaris has been the subject of agricultural extension activity in northern Australia

since the 1920s (Humphries 1967). In the 1950s C. ciliaris became the prominent sown

pasture grass for the more arid zones of northern Australia and was well researched for its

potential to improve pastures across Queensland, Western Australia and the Northern

Territory (Hall 2001). Most plantings have taken place since the late 1950s (Paull and Lee

1978). It has also been used for soil stabilization and erosion control (Albrecht and Pitts

1997; Grigg et al. 2000), although in some situations it fails to provide effective surface

cover (Harwood et al. 1999).

There have been 580 direct official Cenchrus accessions introduced to Australia from 35

countries (Hall 2001). Nine cultivars are described in the Register of Australian Herbage

Plant Cultivars (Oram 1990). These are usually categorized on the basis of height. Tall

varieties (up to 1.7 m) include ‘Biloela’ and ‘Nunbank’, medium height varieties (up to 1 m)

include ‘Gayndah’ and ‘American’ and short varieties (up to 0.4 m) include ‘West

Australian’ (Paull and Lee 1978). ‘Gayndah’, ‘American’ and ‘Biloela’ are the most widely

sown varieties (Cavaye 1991).

Today C. ciliaris is well established in arid and semi-arid zones in Western Australia, South

Australia, New South Wales and in the MacDonnell Ranges bioregion, Northern Territory

(Pitts and Albrecht 2000; Franks and Hannah 2001). It is by far the most widely distributed,

sown pasture grass in Queensland (Cavaya 1991) and is estimated to be abundant or

dominant over 30 to 50 million ha (Woinarski 2001). In north Queensland, C. ciliaris forms

extensive stands in the Desert Uplands bioregion (Sattler and Williams 1999) and is a

prevalent species in the Dalrymple Shire (Roger Lawes pers. comm.).

30

2.1.6.2. Physical description and growth characteristics

Cenchrus ciliaris is a summer growing, perennial, C4 tussock grass. It grows up to 1.7 m tall,

depending on variety and growing conditions, and has a large, deep root system with some

cultivars having short rhizomes (Paull and Lee 1978). Its stems are erect to somewhat

prostrate, often kneed towards the base, and are extensively branched (Cunningham et al.

1981). Leaves are basal and cauline, with blades 3-25 cm long and 4-10 mm wide (Sharp and

Simon 2002). It has a cylindrical ‘foxtail’ seed head, 2.5-15 cm long and 8-16 mm in

diameter. The rachis of the seed head is a serrated stalk to which clusters of one to three

spikelets are attached by very short stalks. Each cluster, or fascicle, is surrounded by a cup-

shaped circle of bristles (involucre) (Paull and Lee 1978) and may contain one to five seeds

(Humphreys 1981). Cenchrus ciliaris produces seed apomictically so cultivars breed true to

type. However, odd plants do reproduce sexually and these have been used for cultivar

development (Paull and Lee 1978).

Cenchrus ciliaris is grown in regions receiving 300 to 1000 mm average annual rainfall

(Humphreys and Partridge 1995). It is less productive than other species in higher rainfall,

coastal areas and its growth is checked by frost (Paull and Lee 1978). Its drought and grazing

tolerance are associated with its deep spreading root system and its characteristic of stem bud

development being slightly below ground level (Marriott 1955). It also has swollen stem

bases that accumulate carbohydrates, allowing it to survive drought and fire, and to green up

more rapidly than other species after rain (Humphreys 1974). Cenchrus ciliaris does not

have day length requirements for flowering, which is promoted by rain (Humphreys and

Partridge 1995). There is a long period of head production during the growing season

(Humphreys 1981) and it produces abundant seed (Franks et al. 2000; Hall 2001).

Cenchrus ciliaris is adaptable in its soil requirements. It grows in soils of moderate fertility

but with variable textures, preferring lighter textured soils but it still grows well on self-

mulching soils (Humphreys 1974). Although it is not as demanding of nutrients as some

other introduced species (Humphreys and Partridge 1995), establishment is quicker and

drought tolerance superior on high phosphorus soils (Paull and Lee 1978). It has only

moderate salt tolerance and is sensitive to water logging (Anderson 1972; Humphreys 1974).

Cenchrus ciliaris is recognised as strongly competitive once established. However, its spread

and colonising ability is more contentious (McIvor 2003). Spread of C. ciliaris from

established pastures has often been low or non-existent and is less than for some other

introduced grasses (Hacker 1989; McIvor 2003). In contrast, as an invasive weed C. ciliaris

31

has been reported to be a highly successful invader and its ease of spread and invasive nature

is cited as a serious problem (Latz 1991; Griffin 1993). The spread of C. ciliaris appears

related to the degree to which disturbance reduces competition from existing flora

(Humphries 1976; Cavaya 1991). McIvor (2003) pointed out that competition from resident

plants is critical to controlling the spread of C. ciliaris: it is able to colonise bare areas but is

unable to invade dense vegetation.

2.1.6.3. Cenchrus ciliaris – valuable pasture species or destructive invader?

In a pastoral context, C. ciliaris is highly valued for its wide adaptation, productivity under

grazing and, particularly, its drought tolerance (Hall 2001). It “has converted low producing

shrublands to highly productive grazing lands” (Paull and Lee 1978) and “brought great

financial benefit to many individual producers and companies” (Hall 2001). It is well

adapted to a wide range of soils and the climate of arid and semi-arid Australia (Hall 2001)

and has become one of the most important pasture introductions for this region. In fact, in

many inland areas it is the only suitable sown grass (Cavaya 1991).

However, while the spread and persistence of C. ciliaris have long been valued by the

pastoral industry, its domination of native vegetation and the associated changes in species

richness and fire regimes are of concern to non-pastoral land managers, particularly those

responsible for conservation areas. Low (1997) considered C. ciliaris to be “perhaps the

most destructive” of Australia’s introduced pasture grasses. It has been associated with lower

plant diversity both in Australia (Fairfax and Fensham 2000; Franks 2002) and overseas

(Saucedo-Monarque et al. 1997; Daehler and Carino 1998). It displaces native vegetation,

for example Triodia spp. (Spinifex) and Aboriginal food plants in Western Australia

(Keighery 1991). It may also affect other biota and has been associated with reduced

invertebrate (Best 1998) and vertebrate (QPWS 2001) diversity. In central Queensland,

Ludwig et al. (2000) found that Delicate Mouse (Pseudomys delicatulus) numbers declined

as C. ciliaris cover increased. Cenchrus ciliaris is an aggressive coloniser of moist habitats,

such as river levees and alluvial pans, where it forms dense monocultures (Humphries et al.

1991). It is a major threat to key mesic areas in the arid zone where it has spread rapidly

during periods of high rainfall and flooding (Griffin 1993). These mesic habitats are critical

parts of the landscape where rare and relic native species occur and colonization of C.

ciliaris in these sites is believed to threaten the survival of these species (Humphries et al.

1991).

32

Cenchrus ciliaris is also considered to be altering local fire regimes. It produced more

biomass and cures later than the native perennial grasses, causing hotter fires later in the

season and an increased incidence of fire (Humphries et al. 1991; Latz 1991). It is a threat to

dry rainforest remnants as it grows along the edges and within the remnants, carrying hot

fires that progressively destroy the rainforest (Fensham 1996). It appears that there is

positive feedback between C. ciliaris and fire. Cenchrus ciliaris increases the risk of fire and

the abundance of C. ciliaris has been found to increase after fire (Butler and Fairfax 2003). It

has also been suggested that C. ciliaris has allelopathic effects (Cheam 1984ab; Nurdin and

Fulbright 1990).

Many Cenchrus species have traditionally been considered weeds as their burrs contaminate

wool and their barbed spines damage skin (Harden 1993). As early as 1906 C. australis R.

BR. (Australian) (hillside burr grass) was included in “The weeds and suspected poisonous

plants of Queensland” (Bailey 1909). Today the environmental impacts of C. ciliaris in

particular are recognized and it is included in various weed lists for Australia (e.g. the

“CSIRO’s handbook of Australian weeds”, Lazarides et al. 1997) and overseas (e.g. Hawaii,

Smith 1985). Although acknowledged as one of Australia’s worst environmental weed

(Humphries et al. 1991), C. ciliaris is not currently listed as a weed of national significance

(www.weeds.org.au/natsig.htm).

Despite the assertion, in conservation management circles at least, that C. ciliaris has

negative impacts on biodiversity and fire regimes, there is little evidence of these impacts, in

terms of quantitative data, in the literature. The impacts of C. ciliaris are often stated without

supporting data. For example, the report by Humphries et al. (1991) is a commonly cited

reference to C. ciliaris impacts. However, no data are provided in this report and the authors

acknowledge that the information given is anecdotal. Although some published studies do

provide quantitative data, these generally report comparisons between invaded and non-

invaded areas (e.g. Fairfax and Fensham 2000). There are difficulties in determining the

impacts of invasive species using this approach since the data are essentially correlative and

do not provide evidence of cause and effect (Wheeler and Giller 1982). Consequently, in

addition to the need for more quantitative data describing this species in an ecological

context, studies are required to investigate the mechanisms underlying C. ciliaris effects.

Determining the impacts of invasive plants is problematic. Issues relating to investigating

invasive plant effects and techniques available are briefly discussed below.

33

2.1.7. MEASUREMENT AND ASSESSMENT OF INVASIVE PLANT IMPACTS

Determining the impacts of invasive plants is problematic on three fronts. Firstly, what

should be measured? Invasive plant impacts may vary depending on what is measured and at

what scale it is measured. Secondly, measurement of plant impacts is often difficult

technically, complicated by the fact that pre-invasion data are often not available and

invasive plant effects are often confounded with disturbance effects. The third issue is how

to assess any measurement of impact.

2.1.7.1. What to measure

Invasions may be deemed positive or negative depending on what is investigated. For

example, the invasive tree Tamarix aphylla has been associated with lower numbers of

reptiles and most birds, but it has a positive effect on numbers of aerial, insectivorous birds

(Griffin et al. 1989). D’Antonio et al. (2000) noted that their data suggested that the

“impacts of exotic species on community composition are both species and context

dependent”. In their study of the impacts of exotic grasses on native plant composition in

Hawaii they found that not all grasses had the same impact, the impact of a particular species

varied over the range in which it was found, and the species varied in their ranges.

The population dynamics of the invader and response of the community vary over space and

time. Therefore, in addition to determining what should be measured, the spatial and

temporal scales of measurement are important (Parker et al. 1999). For example, McIvor

(1998) reported oversowing native pastures with a mixture of exotic legumes and grasses

reduced the total number of species recorded on a plot basis but increased the numbers found

at a smaller scale. The detected effects of invasive species may vary considerably depending

on the timing of measurements also. For example, the cover of perennial grasses and annual

plants is reduced under newly established Larrea tridentata shrubs in New Mexico.

However, the highest cover of perennial grasses is found under their canopies once they are

mature (Whitford et al. 2001).

Adair and Groves (1998) recommended measuring a broad range of environmental

parameters to determine invasive plant impacts on diversity, ecosystem-level functions and

successional consequences. However, while some properties, such as plant productivity may

be relatively easily measured others, such as ‘stability’ may be very difficult to measure

(Walker and Smith 1997).

34

2.1.7.2. How to measure plant impacts

Four principal techniques can be used to determine invasive plant impacts: (1) multi site

comparisons, (2) weed removal, (3) weed addition, and (4) time series studies. These are

reviewed by Adair and Groves (1998) and Walker and Smith (1997).

The multi site comparison approach involves comparing sites where the invader species is

present with sites where the invader is absent (Adair and Groves 1998). The advantage of

this method is that it allows detailed data to be collected within a short time period.

However, considerable care needs to be taken in matching invaded and control sites and to

explore alternative hypotheses that could explain any observed differences. The method

assumes that the invaded and control sites were similar prior to the invasion (Adair and

Groves 1998) but very often the relevant pre-invasion data needed to verify this are not

available. Invasion often occurs simultaneously with habitat modification (D’Antonio et al.

1998), making it difficult to differentiate between effects of the plant and disturbance

(Vitousek 1986; Woods 1997; Lambrinos 2000). Invasive species are often so widespread

that by the time they are noticed it is impossible to identify comparable control (uninvaded)

and invaded sites for study (D’Antonio et al. 1998).

Weed removal and addition studies use manipulative techniques to investigate the impacts of

invasive plants. Weed removal studies, where the invasive species is totally or partially

removed, may provide strong evidence of invasive plant impacts (Adair and Groves 1998).

However, the removal process may itself impact on the system, thereby confounding

disturbance and invasive plant effects (D’Antonio et al. 1998). Plant removal studies have

been criticized because of potential problems such as soil disturbance during removal, soil

compaction during monitoring and unknown effects of leaving root material in the soil (refer

D’Antonio et al. 1998). A further limitation of this method is that a long time frame may be

required to detect changes due to plant removal. This method is not suitable in situations

where the invader has caused irreversible damage such that the ecosystem is unable to return

to its uninvaded state (Walker and Smith 1997; Adair and Groves 1998).

Weed addition studies can provide irrefutable evidence of the impacts of invasive plants as

pre-invasion conditions can be measured. However, long monitoring times may be required

to assess any impacts (Adair and Groves 1998). The greatest disadvantage with this method

is the issue of weeds escaping and invading new areas. As noted by Vermeij (1996),

“Invaders can have unforeseen and often destructive effects on recipient communities. The

field of invasion biology should therefore adopt the standard that work with non-native

35

species or populations be carried out under strictly controlled laboratory conditions from

which accidental release of individuals, including gametes and dispersal stages is

impossible”. Weed addition experiments are commonly conducted under artificial

conditions, for example in pots in a glasshouse, and very few attempts have been made to

use this technique under field conditions (Adair and Groves 1998).

Time sequence studies involve monitoring the impact of invasive species at a site over time

(Adair and Groves 1998). Walker and Smith (1997) stated that the best way to measure the

impact of an invader is to have measurements before, during and after invasion has occurred.

However, such opportunities are rare and again, this technique may require long monitoring

periods to detect invasive plant impacts.

As the limitations of these different techniques indicate, the measurement of invasive plant

impacts is hindered by a number of factors. Lack of detailed background information and the

coarseness of most ecosystem-level measurements make it difficult to detect small or subtle

effects (Vitousek 1990). For example, invasive plant species are likely to alter plant-

herbivore interactions but the wide fluctuations in herbivore populations, particularly

following a disturbance, make it difficult to determine the effects of an invasive plant

(Breytenbach 1986). Most studies documenting changes in species diversity or vigour have

used the multi-site comparison technique (Adair and Groves 1998). This is essentially a

survey technique and its value in determining invasive plant effects is limited by problems of

lack of pre-invasion information and confounding factors as described above. It provides

only correlative data and lacks the power of manipulative studies for determining causal

relationships (Adair and Groves 1998). Consequently, there is a need for manipulative

experiments in invasive plant impact studies (Walker and Smith 1997). However, the more

powerful, controlled, replicated methods are the most difficult to conduct (Woods 1997;

Parker et al. 1999).

2.1.7.3. Impact assessment

The assessment of invasive plant impacts is a critical one for management. If the impact of

an invasive plant is not quantitatively assessed, valuable resources may be wasted in control

programs (Adair and Groves 1998). In Australia, there are no uniformly recognised national

criteria to assess environmental weed effects (Humphries et al. 1991). Assessing the impacts

of invasive plants is not straight forward. In contrast to invasions on agricultural land where

impacts of invasive plants can be assessed in terms of economic losses, impacts of invasive

36

plants on natural and semi-natural systems involve some kind of ‘ecological damage’ that

can be hard to define (Eser 1998) and evaluate in economic terms. As noted by Williams and

West (2000), problems caused by environmental weeds cannot be simply stated in economic

terms because they include issues of ecosystem stability, function and biodiversity.

Central to assessing invasive plant impacts is determining what level of change will be

considered an impact and how significant that impact is (Morrison 1997). Parker et al.

(1999) noted that relatively little attention has been placed on developing generalizations

regarding the level of invasive species impact. They pointed out that there can be surprising

disagreement over the magnitude of invasive species impacts, due, in part, to lack of baseline

data on the original ecosystem structure or function as well as the fact that there is no

common framework for quantifying or comparing the total impacts of invaders. They

proposed describing the overall impact of an invasive species in terms of the total area

occupied, the abundance of the species and some measure of the impact per individual.

However, they acknowledged that while measuring area and abundance may be straight

forward, quantification of ecological effects is not.

2.1.8. CONCLUSIONS

Invasive plants are of increasing ecological importance world wide (Groves 1991) and exotic

grasses in particular have been involved in some of the most destructive and widespread

invasions (Parsons 1972; D’Antonio and Vitousek 1992). As superior competitors for light,

water and nutrients, invasive, exotic grasses often dominate vegetation and displace native

species and grass invasions have been associated with reduced biodiversity and the alteration

of successional processes (D’Antonio and Vitousek 1992). As well as directly affecting

vegetation composition and structure, exotic grasses may cause major habitat change via

effects on ecosystem processes. Their effects on fire regimes are probably the most

significant. Exotic grasses often produce more biomass and burn later in the season than the

native species they replace (Low 1997). They may promote hotter, more frequent fires and

introduce fire into areas where it was previously rare or absent (Smith 1985; Mack and

D’Antonio 1998). The resultant changes to fire intensity and timing may have profound

implications for ecosystem structure and function (Macdonald et al. 1989; Vitousek 1990).

Such major, habitat-changing impacts mean that exotic grasses are a serious problem in

many different ecosystems worldwide.

37

Until recently, the ecological impacts of grasses in natural ecosystems in Australia had

received little attention (Adair and Groves 1998). However, today there is considerable

concern about their negative impacts and particular attention has been drawn to those species

deliberately introduced for pasture improvement (Humphries et al. 1991; Lonsdale 1994;

Low 1997; Whitehead and Wilson 2000). Invasive, introduced grasses are believed to have a

profound effect on a broad range of plant and animal communities in Australia at both the

habitat and landscape level (Adair and Groves 1998). The problems associated with

introduced grasses were highlighted by Humphries et al. (1991) in their report on

environmental weeds in Australia. Of the 18 species they listed as Australia’s worst

environmental weeds, five are tropical pasture grasses. One of the most controversial

inclusions in this list is C. ciliaris.

In a pastoral context C. ciliaris has been highly regarded, both for its value as pasture for

livestock and for its soil protecting properties (Hall 2001). However, more recently it has

received increasing attention because of its apparent effects on biodiversity and fire regimes

(Humphries et al. 1991; Low 1997). Cenchrus ciliaris is a strong competitor and its high

seed yields and light, fluffy seed allow it to spread readily via wind and water (Hall 2001). It

aggressively colonizes moist habitats, such as river levees, and has been reported to form

dense monocultures displacing native vegetation (Humphries et al. 1991). Its effects on

native flora may flow-on to affect native animals (Humphries et al. 1991; Ludwig et al.

2000) and C. ciliaris has been associated with decreased plant (Fairfax and Fensham 2000;

Franks 2002), invertebrate (Best 1998) and vertebrate (QPWS 2001) diversity. In addition to

directly out-competing native species, C. ciliaris invasion may result in major habitat change

via its effects on fire regimes. Cenchrus ciliaris invasion leads to hotter late-season fires and

an increased incidence of fire (Humphries et al. 1991; Latz 1991; Butler and Fairfax 2003).

Consequently, invasion by C. ciliaris is seen as a major threat to key mesic habitats in the

arid zones (Humphries et al. 1991) and fire sensitive vegetation (Woinarski 2001).

Although C. ciliaris is now considered one of Australia’s worst environmental weeds (State

of the Environment Advisory Council 1996), there are relatively few published studies

quantifying its ecological effects. Given the extensive distribution of C. ciliaris in

Queensland, its potential for further spread, and the controversy regarding its ‘value’, the

ecological impacts associated with this species need to be determined.

38

2.2. IS THERE A RELATIONSHIP BETWEEN HERBACEOUS SPECIES

RICHNESS AND CENCHRUS CILIARIS ABUNDANCE?

2.2.1. INTRODUCTION

There are few published studies quantifying relationships between C. ciliaris and species

richness. Cenchrus ciliaris has been associated with reduced plant species richness in central

Queensland (Fairfax and Fensham 2000; Franks 2002) whereas in north-eastern Queensland,

both increases and decreases in species richness in C. ciliaris-dominated vegetation were

found at one site, depending on the scale of measurement (McIvor 1998).

In this chapter two studies are described. The overall aim of the studies was to document

herbaceous species richness patterns in relation to C. ciliaris. In the first study, the

relationship between C. ciliaris and herbaceous species richness was investigated by

comparing species richness of C. ciliaris-dominated sites and non-C. ciliaris sites in the

Dalrymple Shire, north-eastern Queensland. In such site-comparison studies it is often

difficult to determine the effects of exotic plants on species richness since other factors, such

as soil fertility and grazing regime, may influence species richness and their effects are often

confounded with those of the invader. In the second study, the relationship between C.

ciliaris and species richness was investigated at a single site in the absence of grazing to help

minimize some of these confounding factors. Rather than comparing species richness

between areas with and without C. ciliaris, in this study the relationship between C. ciliaris

and herbaceous species richness was investigated by comparing areas with varying levels of

C. ciliaris biomass.

There is a considerable literature dealing with the relationship between plant species richness

and community productivity, often measured as biomass (see reviews by Grace 1999; Waide

et al. 1999; Mittelbach et al. 2001). It has been suggested that there is a general relationship,

described by the well known ‘humped-back curve’ (Grime 1973; 1979), in which species

richness is highest at intermediate biomass levels and lower at high and low biomass,

reflecting changes in stress and/or disturbance levels. However positive, negative and no

relationship between biomass and species richness have also been reported (Waide et al.

1999). The aim of the second study was to determine if there was a relationship between

herbaceous species richness and C. ciliaris biomass and whether the species richness-

biomass relationship for C. ciliaris was different from that for a dominant native species.

Herbaceous species richness and biomass data were collected from plots varying in

39

composition from dominance by C. ciliaris to dominance by Bothriochloa ewartiana, a

native perennial grass. This enabled comparison of patterns of species richness in relation to

C. ciliaris and B. ewartiana biomass.

2.2.2. METHODS

2.2.2.1. Study area

Sites for both studies were located within the Dalrymple Shire, north-eastern Queensland,

Australia. The Dalrymple Shire covers an area of 68,850 km2, extending from 22o05’S to

18o30’S and is bounded by the Great Dividing Range on the west and a chain of coastal

ranges on the east. The climate is dry tropical with an average annual rainfall of between 500

and 700 mm, 80% of which falls between December and April. However, rainfall is

extremely variable from year to year (Quirk et al. 1997; Ash et al. 2002). There are many

geological landscapes in the region, which give rise to a complex mixture of land types (Ash

et al. 2002). Soils vary from sands and massive earths to cracking clays and are generally

low in nitrogen and phosphorus. The main vegetation type is eucalypt woodland with a

grassy understorey. There are scattered acacia communities throughout the Shire, open

grasslands in the south-west and some small areas of rainforest, mainly on the coastal ranges

(Quirk et al. 1997).

Livestock grazing is the predominant land use in the Shire (Quirk et al. 1997). Although

native grasses are the main pasture resource supporting the livestock industry, increasing

areas have been sown to introduced legumes and grasses (McIvor 1998). The main sown

exotic grasses are C. ciliaris and Urochloa mosambicensis (sabi grass). As well as deliberate

sowing, significant areas have been colonized by exotic grasses. Cenchrus ciliaris has spread

readily in disturbed river frontage country and other exotic grasses, such as Bothriochloa

pertusa (Indian couch), have become naturalized in the region (Ash et al. 2002).

40

2.2.2.2. Study one: herbaceous species richness with and without Cenchrus ciliaris

Study sites

Herbaceous species richness was surveyed at seven C. ciliaris-dominated sites and nine non-

C. ciliaris sites between April and June 2000. Sites were selected on the basis of dominant

herbaceous species, being either dominated (in terms of biomass) by C. ciliaris or not. It was

not possible to find paired sites (neighbouring sites with and without C. ciliaris) due to

limited resources. The sites used here were selected on the basis that they were accessible

given the resources available. Sites included a range of landscape types and grazing regimes

that varied from currently ungrazed to grazed by livestock and native marsupials (Table 2.2).

At each site, a survey plot (8 m by 8 m) was positioned to avoid trees, shrubs, grazed-out

patches and scalds. Where possible, plots without C. ciliaris (non-C. ciliaris plots) were

positioned to avoid large patches of exotic grasses. However, two non-C. ciliaris plots were

dominated by B. pertusa.

Table 2.2. Latitude and longitude, soil type and current grazing regime of C. ciliaris and

non-C. ciliaris plots surveyed in the Dalrymple Shire.

Latitude Longitude Soil type* Current grazing

Cenchrus ciliaris plots 20o13´S 146o34É Chromosol yes 20o07´S 146o19É Chromosol yes 20o10´S 146o30É Chromosol yes 19o25´S 145o51É Tenosol yes 19o49´S 146o06É Tenosol no 19o53´S 146o11É Ferrosol yes 19o41´S 145o45É Ferrosol no Non-C. ciliaris plots 20o12´S 146o34É Chromosol no 20o27´S 145o44É Kandosol no 20o21´S 145o48É Kandosol yes 20o10´S 146o28É Chromosol yes 20o14´S 146o40É Chromosol yes 20o33´S 146o08É Sodosol yes 19o53´S 146o11É Ferrosol yes 19o53´S 146o11É Ferrosol yes 19o41´S 145o46É Ferrosol no

*Soil type determined using the Land Resources of the Dalrymple Shire (Rogers et al. 1999)

data base. Soils were classified to order using the Australian Soil Classification (Isbell 1996).

41

Vegetation sampling

In each plot, herbaceous species were recorded using a nested plot design (Rosenzweig

1995). Given the importance of scale in assessing patterns of species richness (Bond and van

Wilgen 1996), this method was chosen to provide data at a range of scales. Starting at one

corner, data were collected from areas doubling in size from 1-64 m2, providing seven scales.

A string grid was laid out over each plot to facilitate data collection. Species were counted

by recording the species found at the first scale and then recording any new species found at

each successive scale. While successive scales doubled in size, all scales were searched on a

1 m2 basis using a square metal frame moved over the area to outline contiguous quadrats.

When determining species numbers, unidentifiable plants were counted if they were distinct

from species already counted. Unknown species were collected for later identification. Plants

that could not be assessed as different, such as small seedlings, were excluded from the

species counts. Numbers of species of legumes, other forbs (referred to from here on as

forbs), sedges, perennial grasses, annual grasses, exotic and ‘rare’ species were determined.

Rare species were defined as those species found in only one plot in the survey. Numbers of

species in some plant groups were under-estimated since unidentified plants could not be

categorized.

Species-area curves

Species-area curves for C. ciliaris and non-C. ciliaris plots were derived using the power

function in CurveExpert 1.3 (Hyams 2003). Species-area curves relate species number and

area, usually by the power function: S = cAz where S is the number of species found in area A

(Williams 1996). C and z are estimable parameters, c being the expected number of species

in a unit areas and z being the instantaneous rate by which species richness increases with an

incremental increase in area (Pastor et al. 1996).

2.2.2.3. Study two: herbaceous species richness with varying Cenchrus ciliaris biomass

Study site

The second study was conducted at Hillgrove Station (19o40’S; 145o45’E). The soil is a

eutrophic red ferrosol (Isbell 1996) developed on basalt and is of moderate fertility. The

vegetation is open woodland with an average tree density of 64 trees/ha and average tree

basal area of 5.5 m2/ha. The upper stratum is comprised mainly of Eucalyptus crebra

42

(ironbark) and Corymbia erythrophloia (bloodwood) trees (McIvor et al. 1991). The

herbaceous layer is dominated by perennial grasses including C. ciliaris and native species

such as B. ewartiana, Heteropogon contortus and Chrysopogon fallax. An experimental site

had been established on this property for a different study in 1981 and a detailed site

description is given by McIvor et al. (1991).

The area used for this study was a spare paddock within the original experimental site. This

paddock, of approximately 3 ha, was used to hold cattle from time to time and has been

ungrazed or very lightly grazed since 1981. The herbaceous layer was dominated by two

perennial grasses, C. ciliaris and B. ewartiana. Cenchrus ciliaris was not originally present

at the site (McIvor et al. 1991). It was sown in some paddocks along with other exotic

species in 1981 and has since spread into unsown areas. Eighteen 8 m by 8 m plots were

located in the paddock in March 2000 using a stratified, random sampling procedure. Plots

were selected to vary in biomass composition, from dominance by C. ciliaris to dominance

by B. ewartiana. Trees and large patches of other grass species were avoided.

Vegetation sampling

In each plot, herbaceous species numbers were recorded using the nested plot design

described above, but in study two data were collected from areas doubling in size from 0.25-

64 m2, providing nine increasing scales/plot. In addition, herbaceous biomass and the

percentage C. ciliaris and B. ewartiana biomass were estimated at each scale. Total

herbaceous biomass was estimated using the BOTANAL technique (Tothill et al. 1992).

This method involves visually ranking quadrats and converting rank scores to biomass using

a relationship determined from a set of ranked and weighed standards. For scales 1-8 m2,

biomass was estimated over the whole area using 1 m2 contiguous quadrats. (Smaller quadrat

sizes were used for the 0.25 and 0.50 m2 scales). Due to time constraints, the whole area of

larger scales (16 m2 and above) could not be surveyed. For these larger scales, the biomass in

additional areas was estimated by surveying half the area using randomly selected quadrats.

For example for the 16 m2 scale, 8 m2 was fully surveyed (as it made up the previous scale).

The biomass of the other 8 m2 was estimated using four randomly positioned 1 m2 quadrats.

In addition to biomass rank, the percentage of C. ciliaris and B. ewartiana biomass present

in each quadrat was estimated to the nearest 5%.

Plots were surveyed in March and April 2000. As it was not possible to rank and cut

standards for each sampling day, a series of 12 permanent biomass standards were set up at

43

the start of the data collection period. The standards were 1 m2 plots ranging from low to

high biomass and varying in dominant species. At the end of each sampling day the

standards were ranked. After all plots had been surveyed the standards were cut at

approximately ground level. Material was considered in the quadrat if it fell within the

vertically projected quadrat boundaries. The material was oven dried at 65oC before being

weighed. For each sampling day, the linear rank–biomass relationship calculated from the

standards was used to convert the biomass ranks of plots surveyed that day.

Data analysis

Relationships between biomass and species richness in study two were investigated using

linear regression. Data from each scale were analysed separately. Different methods were

used to investigate species richness-C. ciliaris biomass relationships at small and large scales

because of differences in biomass composition with scale. At the smallest scales (up to 1

m2), most plots were dominated, in terms of biomass, by either C. ciliaris or B. ewartiana;

that is plot biomass was at least 70% C. ciliaris or B. ewartiana. Linear relationships

between dominant species biomass and species richness for C. ciliaris-dominated plots (n =

9) were compared with those for B. ewartiana-dominated plots (n = 8). At larger scales, plot

biomass composition was more variable and few plots were dominated by one species.

Therefore, biomass-species richness relationships were investigated using all 18 plots

together. Relationships between C. ciliaris biomass and species richness were investigated

using multiple linear regression with C. ciliaris biomass and total biomass as independent

variables (there was little correlation between C. ciliaris and total biomass in these data sets).

This method allowed investigation of C. ciliaris biomass effects on species richness while

allowing for total biomass effects (i.e. total biomass was treated as a covariate).

Relationships between C. ciliaris biomass and the number of grass, forb (here forbs included

one sedge species), legume and forb plus legume species as a group were similarly

investigated. Where there was no total biomass effect, simple linear regression was used.

Relationships between total biomass and species richness and B. ewartiana biomass and

species richness were also investigated using linear regression. Since B. ewartiana biomass

and total biomass were highly positively correlated (average r = 0.724) the multiple linear

regression model was inappropriate.

Correlations between species-area curve parameters and biomass have been used to

investigate species richness-biomass relationships (Pastor et al. 1996; Weiher 1999).

Species-area curves were derived for each plot using CurveExpert 1.3 (Hyams 2003) as

44

described above. Linear regression was used to test for relationships between the parameters

c and z and C. ciliaris biomass and total biomass. All analyses were performed using Genstat

(2001).

2.2.3. RESULTS

2.2.3.1. Study one: herbaceous species richness with and without Cenchrus ciliaris

Mean species richness was lower in C. ciliaris plots than in non-C. ciliaris plots at all scales

investigated (Figure 2.1a). At the smallest scale (1 m2), species richness ranged from 1-4

(mean ± SE of 3 ± 0.4) species/m2 for C. ciliaris plots and from 4-14 (mean ± SE of 9 ± 1.2)

species/m2 for non-C. ciliaris plots. At the largest scale (64 m2), species richness ranged

from 17-32 (mean ± SE of 21 ± 2.0) species/64 m2 for C. ciliaris plots and from 20-53 (mean

± SE of 40 ± 3.7) species/64 m2 for non-C. ciliaris plots.

New species were found in the last sampling area in all plots. On average, the final doubling

of the search area, from 32 m2 to 64 m2, resulted in an additional six species being found in

C. ciliaris plots (a 46% increase in species number) and an additional eight species in non-C.

ciliaris plots (a 25% increase in species number).

Trends in species richness within individual plant groups generally reflected the pattern

found for total species richness. There were fewer forb, legume and perennial grass species

at all scales (Figures 2.1b-d) and fewer sedges at the two largest scales in C. ciliaris plots

compared with non-C. ciliaris plots. Numbers of sedge species at smaller scales and annual

grasses at all scales were too low to determine distribution patterns. In total 47 forb, 26

legume, 28 perennial grass, 13 annual grass and four sedge species were found in the survey

(see appendix 1A).

In terms of species numbers, C. ciliaris and non-C. ciliaris plots differed at all scales.

However, in terms of proportional composition, C. ciliaris and non-C. ciliaris plots differed

at the smallest scales only (Figure 2.2a). At the 1 m2 scale, most species in C. ciliaris plots

were perennial grasses. Only one C. ciliaris plot contained a forb at this scale and only four

legume species were found over all C. ciliaris plots. In comparison, although perennial

grasses were the largest group in non-C. ciliaris plots, forb and legume species were almost

as abundant. All non-C. ciliaris plots contained one or more legumes at the 1 m2 scale and all

45

but two plots contained forbs. This difference in proportional composition declined with

increasing scale with C. ciliaris and non-C. ciliaris plots being similar in terms of

proportional composition by the 4 m2 scale. At the largest scale, forbs were the most

abundant species group in both C. ciliaris and non-C. ciliaris plots (Figure 2.2b).

0

10

20

30

40

0 20 40 60

150

5

10

15

0 20 40 60

0

5

10

15

0 20 40 60

Num

ber o

f spe

cies

0

5

10

0 20 40 60

)

)

)

)

Figure 2.1. Mean (± SE) numbers of

plots (□) at each of seven scales from 1

Figures (b-d) show numbers of non-le

respectively.

Area (m2

(a

(d

(c)

(b

species for C. ciliaris plots (■) and non-C. ciliaris

-64 m2. Figure (a) shows the total number of species.

guminous forb, legume and perennial grass species

46

0

10

20

30

40

50

60

70

80

Forb Legume Sedge Ann.Grass

Per.Grass

Forb Legume Sedge Ann.Grass

Per.Grass

(a) 1 m2 (b) 64 m2

)io

n (%

itpo

som

nal c

opor

tioP

r

Species groups

Figure 2.2. Proportional composition, in terms of numbers of non-leguminous forb, legume,

sedge, annual grass and perennial grass species, of C. ciliaris plots (■) and non-C. ciliaris

plots (□) at (a) 1 m2 and (b) 64 m2 scales.

Exotic species were found in all plots and mean numbers were not very different between C.

ciliaris plots (mean ± SE of 4.6 ± 0.43 exotics/plot) and non-C. ciliaris plots (mean ± SE of

3.7 ± 0.47 exotics/plot). The exotic species identified included four forbs, four legumes, one

annual grass and four perennial grasses. Cenchrus ciliaris plots had fewer native species

(mean ± SE of 9.4 ± 1.39 native species/plot) than non-C. ciliaris plots (mean ± SE of 26.2 ±

2.76 native species /plot).

In this study 28 species were classified as ‘rare’ (occurring in only one plot in the survey).

Most plots contained rare species and there was little difference in mean numbers of rare

species between C. ciliaris plots (mean ± SE of 1.1 ± 0.40 rare species/plot) and non-C.

ciliaris plots (mean ± SE of 2.2 ± 0.52 rare species/plot). Rare species included 14 forbs, six

legumes, one sedge, four annual grasses and three perennial grasses.

The distribution of some individual species/genera differed between plot types. Three forbs,

one legume and five grasses occurred in most non-C. ciliaris plots but were absent from

most or all C. ciliaris plots (Figure 2.3).

47

Taxa

Number of C. ciliaris plots

Number of non-C. ciliaris plots

7 6 5 4 3 2 1 1 2 3 4 5 6 7 8 9 Evolvulus alsinoides Hybanthus enneaspermus Phyllanthus spp. Zornia spp. Aristida spp. Bothriochloa ewartiana Chrysopogon fallax Panicum spp. Tripogon loliiformis

Figure 2.3. Herbaceous taxa showing contrasting distribution between C. ciliaris and non-C.

ciliaris plots. The number of shaded cells indicates the number of plots in which the species

was present.

The mean species-area curves for C. ciliaris and non-C. ciliaris plots differed. While c

values were lower for C. ciliaris plots (mean ± SE of 2.9 ± 0.46) than for non-C. ciliaris

plots (mean ± SE of 9.3 ±1.41), the reverse was true for z values (mean ± SE of 0.51 ± 0.058

and 0.37 ± 0.024 for C. ciliaris and non-C. ciliaris plots respectively).

2.2.3.2. Study two: herbaceous species richness with varying Cenchrus ciliaris biomass

The 1999-2000 wet season experienced higher than average rainfall (866 mm from October

1999 to April 2000), the third consecutive wet season to do so (Ash et al. 2002). These good

seasons and the absence of grazing resulted in very high biomass levels in the Hillgrove

plots. Mean (± SE) estimated biomass at the largest scale (64 m2) was 768 (± 45.0) g/m2. Plot

biomass composition varied as area increased. At the smallest scales (up to 1 m2), all but one

plot were dominated by either C. ciliaris or B. ewartiana: nine plots had greater than 70% C.

ciliaris while eight plots had greater than 70% B. ewartiana. At larger scales, plot

composition was more variable and fewer plots were dominated by a single species (Figure

2.4). Other species generally made a minor contribution to total biomass (for example

between 1 and 29% at the largest scale). At all scales, plots ranged in composition from less

than 1% to greater than 95% C. ciliaris biomass.

48

0

200

400

600

800

1000

1200

16 1 2 4 3 18 17 10 12 14 5 11 9 13 7 15 8 6

Plot identifcation number

Bio

mas

s (g

/m2)

Figure 2.4. Composition of 18 plots in terms of C. ciliaris ( ), B. ewartiana ( ), and other

herbaceous species ( ) biomass (g/m2) at the 64 m2 scale, presented in order of declining C.

ciliaris biomass.

Species richness ranged from 1-4 (mean ± SE of 2.4 ± 0.2) species/0.25 m2 at the smallest

scale (0.25 m2) and from 14-30 (mean ± SE of 20 ± 1.0) species/64 m2 at the largest scale (64

m2). Over all plots 24 forbs, 20 legumes, 17 perennial grasses, three annual grasses and one

sedge were identified (see appendix 1B).

Significant linear trends of declining species richness with increasing C. ciliaris biomass

were found at some scales. At the smallest scales, species richness declined with increasing

C. ciliaris biomass in C. ciliaris-dominated plots and C. ciliaris biomass explained up to

57% of the variation in species number (Figure 2.5). In contrast, no significant linear trends

were found between B. ewartiana biomass and species richness at these scales (P > 0.10) in

B. ewartiana-dominated plots (Figure 2.5).

At intermediate scales also, species richness tended to decline with increasing C. ciliaris

biomass. At these scales, the multiple linear regression model with predictor variables total

biomass and C. ciliaris biomass was significant (P < 0.05) (Table 2.3). After accounting for

the effects of total biomass, C. ciliaris biomass had a significant effect (P < 0.05) on species

richness at the 2 m2 and 8 m2 scales while its effect was not significant at the 0.05 level (P <

0.08) at the 4 m2 and 16 m2 scales. At the two largest scales, the multiple linear regression

49

model was not significant (P > 0.10) and no relationships between species number and C.

ciliaris biomass were detected using simple linear regression (P > 0.10). No significant

relationships between B. ewartiana biomass and herbaceous species richness were found at

intermediate or large scales (P > 0.05) (Table 2.3).

0

2

4

6

8

10

20

S=2.938-0.0004*B P=0.68

S=4.082-0.0027*C P=0.04 Adj R2=0.40

(a)

Num

ber o

f spe

cies

/0.2

5 m

2

0 600 1000 1400

0

2

4

6

8

10

200 600 1000 1400

Num

ber o

f spe

cies

/0.5

0 m

2

S=3.830-0.0008*B P=0.65

S=6.690-0.0055*C P=0.01 Adj R2=0.57

(b)

0

2

4

6

8

10

200 600 1000 1400

Num

ber o

f spe

cies

/1 m

2

S=6.410-0.0027*B P=0.30

S=7.660-0.0059*C P=0.08 Adj R2=0.29

C. ciliaris or B. ewartiana biomass (g/m2)

(c)

Figure 2.5. Herbaceous species number in relation to dominant species (C. ciliaris or B.

ewartiana) biomass (g/m2) for C. ciliaris ( ) and B. ewartiana (□) dominated plots at (a)

0.25 m2, (b) 0.5 m2 and (c) 1 m2 scales. Regression lines and equations are shown. S =

species number, C = C. ciliaris biomass and B = B. ewartiana biomass.

50

Table 2.3. Summary of regression analyses investigating relationships between biomass and

herbaceous species richness at scales 2-64 m2 for (a) C. ciliaris and (b) B. ewartiana.

Multiple linear regressions were significant for scales 2-16 m2 where the predictor variables

were C. ciliaris biomass and total biomass. (The response variable was species number).

(Note: correlation between total and C. ciliaris biomass ranged from r = -0.164 to -0.359).

Simple linear regression was used to investigate C. ciliaris biomass-species number

relationships at 32 and 64 m2 scales (since total biomass was not significant) and B.

ewartiana biomass-species number relationships at 2-64 m2 scales; multiple linear regression

being inappropriate because B. ewartiana and total biomass were highly correlated.

Scale Parameter Estimate SE t prob. Regression F. prob. Adj. R2*

2 m2 (a) Constant 9.38 1.48 <0.001 0.01 0.41 C. ciliaris biomass -0.0038 0.0014 0.014 Total biomass -0.0047 0.0017 0.013 (b) Constant 4.42 0.79 <0.001 0.75 B. ewartiana biomass 0.0004 0.0013 0.745 4 m2 (a) Constant 12.47 2.20 <0.001 0.02 0.31 C. ciliaris biomass -0.0038 0.0020 0.077 Total biomass -0.0072 0.0027 0.016 (b) Constant 6.32 1.00 <0.001 0.73 B. ewartiana biomass -0.0006 0.0019 0.734 8 m2 (a) Constant 18.30 3.02 <0.001 0.01 0.41 C. ciliaris biomass -0.0077 0.0026 0.010 Total biomass -0.0105 0.0036 0.011 (b) Constant 8.14 1.28 <0.001 0.85 B. ewartiana biomass 0.0050 0.0025 0.847 16 m2 (a) Constant 22.87 4.00 <0.001 0.03 0.30 C. ciliaris biomass -0.0076 0.0037 0.058 Total biomass -0.0131 0.0047 0.013 (b) Constant 11.84 1.52 <0.001 0.55 B. ewartiana biomass -0.0018 0.0029 0.547 32 m2 (a) Constant 14.96 1.54 <0.001 0.54 C. ciliaris biomass 0.0029 0.0046 0.537 (b) Constant 17.54 1.62 <0.001 0.19 B. ewartiana biomass -0.0044 0.0032 0.194 64 m2 (a) Constant 19.27 1.65 <0.001 0.44 C. ciliaris biomass 0.0040 0.0050 0.440 (b) Constant 22.72 1.76 <0.001 0.11 B. ewartiana biomass -0.0056 0.0034 0.114 *Adj. R2 is the difference between residual and total mean squares expressed as a proportion of the total mean square (Lane and Payne 1998). Herbaceous species richness tended to decline with increasing total biomass, but only at

51

intermediate scales (simple linear regression P = 0.06, 0.04, 0.09 and 0.06 for scales 2, 4, 8

and 16 m2 respectively). However, the amount of variation in species number explained by

total biomass was less than 20%. (Other parameters affecting species number were not

investigated). No linear relationships between species number and total biomass were found

at the larger scales.

Significant relationships between C. ciliaris biomass and numbers of species in individual

plant groups were found at some scales (P < 0.05) (Table 2.4). However, it is not clear

whether any group was more responsive to C. ciliaris biomass than others. Trends of

decreasing species numbers with increasing C. ciliaris biomass were found for grasses,

legumes and forbs plus legumes. No relationship between forb species and C. ciliaris was

detected except at the largest scale at which forb species richness increased with increasing

C. ciliaris biomass. No relationships between numbers of species in individual plant groups

and B. ewartiana biomass were found except at the largest scale where the numbers of forb

and forb plus legume species declined with increasing B. ewartiana biomass (P < 0.05).

The species-area curve parameters z and c were not linearly related (P > 0.05) to C. ciliaris

biomass or to total biomass.

Table 2.4. Summary of regression analyses investigating relationships between C. ciliaris

biomass and the numbers of species of forbs, legumes, forbs plus legumes and grasses at

scales from 2-64 m2. Values are t probabilities for the estimates of C. ciliaris biomass where

significant (P < 0.05). (Non-significant t probabilities are denoted ‘ns’). Multiple linear

regression, using C. ciliaris and total biomass as predictor variables, was applied when the

total biomass effect was significant (P < 0.05) (denoted *). Otherwise simple linear

regression was used. The direction of statistically significant trends between C. ciliaris and

species number is given in brackets.

Scale Forbs Legumes Forbs+Legumes Grasses 2 m2 ns ns 0.017 (-)* ns 4 m2 ns* ns ns* ns 8 m2 ns 0.014(-)* 0.023 (-)* 0.017 (-)* 16 m2 ns 0.019 (-)* 0.050 (-)* ns 32 m2 ns ns ns ns 64 m2 0.023 (+) ns ns ns

52

2.2.4. DISCUSSION

Cenchrus ciliaris has been associated with decreased plant species richness both in Australia

(McIvor 1998; Fairfax and Fensham 2000; Franks 2002) and overseas (Saucedo-Monarque

et al. 1997; Daehler and Carino 1998). The two studies reported here also provide

quantitative data demonstrating a negative association. In study one, species richness of C.

ciliaris plots was, on average, only 53% that of non-C. ciliaris plots. This difference falls

within the range found by Fairfax and Fensham (2000) who reported that the species

richness of C. ciliaris pastures was 29% of that of native pastures in cleared brigalow

woodlands and 65% of that of native pastures in cleared eucalypt woodlands in central

Queensland. The trend of decreasing herbaceous species richness with increasing C. ciliaris

biomass found in study two provides further evidence of this negative association.

Species richness was lower in C. ciliaris plots than in non-C. ciliaris plots at all scales

investigated. These findings contrast with those of McIvor (1998) who reported that pastures

oversown with C. ciliaris had higher species richness compared with native pastures at the

quadrat scale (0.25 m2) but lower species richness at the plot scale (12.5 m2). Patterns of

species richness may vary across spatial scales (Bond and van Wilgen 1996) and, although

C. ciliaris was found to be associated with reduced species richness at the scales

investigated, it is not clear how the species richness of C. ciliaris-dominated vegetation

compares with that of non-C. ciliaris vegetation at much larger scales. The species-area

curves predict that C. ciliaris plots will have more species than non-C. ciliaris plots at large

scales since the mean z for C. ciliaris species-area curves was greater than that of non-C.

ciliaris curves. However, caution is required in extrapolating species-area curves. The power

function model generally applied to the species-area relationship may not be the best model,

or even a valid model, for fitting data at large scales (Kilburn 1966; Williams M. R. 1995).

The power function is unbounded and so predicts that as area increases, species increase

without limit. However, it is known that species-area curves are asymptotic. In addition, in

this study it appears that the plot area was not large enough to adequately sample community

species richness. In all plots, a considerable number of additional species were found in the

last search area and additional species were frequently observed outside the plot boundaries.

Consequently, c and z have been derived from only a portion of the curves and the shapes of

these curves, and estimates of c and z, are likely to alter as larger areas are added. Given

these constraints, it is not valid to extrapolate these species area curves to predict species

richness patterns at much larger scales.

Although no absolute conclusions can be drawn from the species-area curves, there is some

53

evidence that species richness differences between C. ciliaris-dominated vegetation and non-

C. ciliaris vegetation may diminish at larger scales. The magnitude of the difference between

the two vegetation types decreased with increasing scale. The species richness of non-C.

ciliaris plots was 3.5 times that of C. ciliaris plots at the 1 m2 scale but only 1.9 times higher

at the 64 m2 scale. Also, C. ciliaris and non-C. ciliaris plots were more similar in terms of

proportional composition as scale increased. Further work is required to determine species

richness patterns in relation to C. ciliaris dominance at larger scales.

Both the spatial and temporal scales at which measurements are made are important

considerations in ecological studies (Critchley and Poulton 1998; MacNally and Quinn

1998). Patterns of species richness at small spatial scales may not reflect species richness

patterns at larger scales. Temporal changes in species richness patterns are also difficult to

predict as the influences of disturbance and competition on species richness change over

time. The role of disturbance may be detrimental or beneficial depending on its

characteristics and those of the vegetation. For example, increases in C. ciliaris biomass may

alter local fire regimes (Humphries et al. 1991; Latz 1991; Butler and Fairfax 2003) resulting

in the loss of some native species. Alternatively, disturbances including fire may prevent the

competitive exclusion of other species by C. ciliaris by creating gaps in the sward.

The negative association between C. ciliaris and species richness found in study one was

supported by the results of study two in which negative C. ciliaris biomass-species richness

relationships were found. Species richness may be influenced by site biomass and many

studies report biomass-species richness relationships (e.g. Wheeler and Giller 1982; Moore

and Keddy 1989; Wisheu and Keddy 1994). However, the effects observed here were not

simply total biomass effects: C. ciliaris biomass effects were detected in addition to total

biomass effects. Importantly, while total herbaceous species richness responded to C. ciliaris

biomass, it appeared unaffected by B. ewartiana biomass. No relationships between total

herbaceous species richness and B. ewartiana biomass were detected. The negative

association between C. ciliaris biomass and species richness is consistent with the view that

C. ciliaris reduces species richness. The results also support those of Franks (2002) who

found a negative correlation between C. ciliaris cover and herbaceous species richness at the

1 m2 and ‘site’ (30 m2) scales in central Queensland. However, the results do not provide an

explanation for the C. ciliaris-species richness association. Explanations for the relationship,

such as competitive and/or allelopathic effects (Cheam 1984ab; Nurbin and Fulbright 1990),

need to be tested experimentally.

Although relationships between C. ciliaris biomass and species richness were statistically

54

significant (P < 0.05), the results should be interpreted with caution because of variability in

the data and inconsistency in results. Occasionally individual plots had high leverage and

their omission alters the outcome of the analyses. In addition, the observed trends were not

statistically significant at all scales within the range over which relationships were detected.

The relationships detected were generally not strong, with biomass explaining less than 50%

of the variation in species richness in most cases.

Determining the role of invasive plants in reducing biodiversity is problematic. Invasion

often occurs simultaneously with habitat modification (Ramakrishnam and Vitousek 1989;

Vitousek 1990; D’Antonio et al. 1998), making it difficult to differentiate between the

effects of the exotic species and other disturbance factors (Vitousek 1986; Woods 1997).

Many studies make comparisons between invaded and non-invaded sites and then imply a

causal relationship between exotic invasion and the differences detected. However,

correlative data do not provide evidence of cause and effect (Wheeler and Giller 1982). Plant

species richness may be influenced by a range of factors that vary across sites and these may

be confounded with exotic plant effects. For example, Fairfax and Fensham (2000)

acknowledged that the precise cause of species loss associated with C. ciliaris pastures in

central Queensland could not be easily identified since grazing and tree clearance could also

have affected species richness. Some of these major confounding factors were eliminated in

the Hillgrove study by investigating the relationship between C. ciliaris biomass and species

richness within a single vegetation type in the absence of grazing. However, the cause of

species richness decline with increasing C. ciliaris biomass remains unclear. Although it is

likely that the observed trends reflect the superior competitive ability of C. ciliaris, it is

possible that species richness is affected by other factor(s) correlated with C. ciliaris

biomass.

The failure to detect consistent, strong biomass-species richness relationships may reflect the

fact that other factors are important in influencing species richness at this site. Although it

was selected for its relative uniformity in terms of soil type and overall vegetation structure,

the site varies in soil nitrogen%, phosphorus% and rockiness (McIvor et al. 1991). These and

other unmeasured factors may be influencing species richness. Another explanation for the

failure to detect consistent, strong biomass-species richness relationships may be that the

range of biomass levels was insufficient. Detecting trends in species numbers is likely to be

more difficult over small ranges in biomass. For example, in their study of wetland plant

species richness–biomass relationships in eastern Canada, Moore and Keddy (1989) detected

species richness trends with biomass only when a wide range of biomass values was

included in the analysis. When more restricted ranges were used, no systematic trends were

55

found. Insufficient variation in independent and dependent variables may make detecting

biomass-species richness patterns difficult (Grace 1999). In my study the range of biomass

levels was greatest at small scales but here the range in species numbers was small. This may

have restricted the capacity to detect patterns.

Variability in both biomass and species number may have been depressed in the favourable

growing conditions experienced during this study. Biomass levels in all plots were very high

compared with average production recorded previously at the site (Ash et al. 2002). The high

biomass produced by exotic grasses contributes to some of the detrimental changes

associated with these plants. For example, C. ciliaris is believed to produce more biomass

than native grasses (Humphries et al. 1991; Butler and Faifax 2003), resulting in changes to

fire regimes. Interestingly, in this study it was the native grass, B. ewartiana, which was

positively correlated with total biomass. Plots dominated by C. ciliaris did not have greater

biomass than plots dominated by B. ewartiana, demonstrating that the presence of C. ciliaris

does not always signify greater biomass. Species richness appeared lower than that found

previously at the site. McIvor (1998) reported means of 4.3 and 3.7 species/0.25 m2

(estimated over eight years) for oversown and native pastures respectively at this site. I

found a mean of 2.4 ± 0.2 species/0.25 m2. At a larger scale, McIvor (1998) reported means

of 28 and 31 species/12.5 m2 (estimated over eight years) for oversown and native pastures

respectively. This compares with only 11 species/12.5 m2 estimated from the mean species-

area curve from my study. It is likely that species richness was depressed in the very high

biomass conditions prevailing during the study. However, this large difference in richness

between the two studies may be due, in part, to differences in data collection methods.

McIvor (1998) collected his data from scattered plots, a method that over-estimates species

numbers relative to using contiguous plots as in the nested design (Rosenzweig 1995).

Species richness appeared unaffected by C. ciliaris or total biomass at scales greater than 16

m2. The species richness-biomass relationship may vary with spatial scale and various

workers have highlighted the importance of scale in understanding species richness-biomass

relationships (Moore and Keddy 1989; Rosenzweig 1995; Waide et al. 1999; Mittelbach et

al. 2001). Moore and Keddy (1989) found a relationship between plant species richness and

biomass when comparing among vegetation types (coarse scale) but were unable to detect

any trends when comparing different areas within vegetation types (fine scale). They

suggested that different patterns and processes are found at different levels of organization.

Pastor et al. (1996) showed that the scale of measurement could influence the shape of the

species richness-biomass relationship and noted that relationships determined at one scale

would not necessarily be the same as those found at another. It is unclear whether the lack of

56

association between species richness and C. ciliaris biomass at larger scales in this study is a

consequence of sampling regime or simply indicates that species richness is unresponsive to

C. ciliaris at larger scales. Plots greater than 16 m2 generally became more alike in terms of

amount and composition of biomass as they exceeded the size of single-species patches.

Sampling larger C. ciliaris patches would help clarify the biomass-species richness

relationships at larger scales.

Relationships between species richness and scale have been summarized using species-area

curves (Rosenzweig 1995) and correlations between biomass and the species-area curve

parameters c and z have been used to demonstrate association between biomass and species

richness (Pastor et al. 1996; Weiher 1999). I found no relationships between C. ciliaris

biomass and c or z estimated from species-area curves derived from data up to 64 m2. Failure

to detect a relationship has been interpreted as evidence that species richness is unresponsive

to biomass (Pastor et al. 1996). Therefore, the results may indicate that biomass is not a

major factor accounting for variation in species richness at this site. Alternatively, the failure

to find relationships between biomass and c and z may be due to the small range of biomass

levels sampled, as explained above, and/or to poor estimation of the parameters. As in study

one, it is likely that the maximum area surveyed was too small to adequately sample

community species richness. Therefore c and z are derived from only a portion of the curve

and are likely to differ from estimates made from more complete curves. Sampling larger C.

ciliaris-dominated areas may provide more accurate parameter estimates.

Although the results from these studies show that C. ciliaris is associated with reduced

species richness, no plant groups appeared more affected than others. In study one, the

association between C. ciliaris and lower species richness was found for all major plant

groups (forbs, legumes and perennial grasses) present. In study two there was no strong

evidence to suggest that C. ciliaris biomass had a greater effect on some groups than others.

Legumes as a group may have been more responsive to C. ciliaris biomass but the results

were not definitive. The increase in forb numbers with increasing C. ciliaris biomass at the

largest scale is puzzling since it contrasted with all other trends found.

No differences in numbers of exotic or rare species were detected between C. ciliaris and

non-C. ciliaris plots in study one. Exotic species richness has been found to be variously

higher or lower in C. ciliaris-dominated vegetation than in non-C. ciliaris vegetation

(Fairfax and Fensham 2000). Relationships between exotic species numbers and C. ciliaris

dominance are difficult to predict. Exotic species richness might be expected to be high in C.

ciliaris-dominated vegetation. Exotic species richness has been shown to be positively

57

associated with soil fertility and disturbance (McIntyre and Lavorel 1994; Stohlgren et al.

1999). Cenchrus ciliaris establishes best on more fertile soils and disturbances promoting C.

ciliaris establishment may favour the establishment of other exotics. Alternatively,

domination by C. ciliaris may competitively exclude other species including other exotic

species. The influence of disturbance and competition on exotic species will vary between

sites and over time. Given the inter-plot variation in fertility, disturbance regimes and time

since C. ciliaris establishment, it is not surprising that an association between C. ciliaris and

exotic species richness was not found. In contrast, native species richness was lower in C.

ciliaris-dominated vegetation as reported elsewhere (McIvor 1998; Fairfax and Fensham

2000). With regard to rare species, these are generally considered to be especially

endangered by exotic invasions (Daehler and Strong 1994) and numbers of rare species have

been found to be negatively correlated with exotic species abundance (McIntyre and Lavorel

1994). In this study the numbers of rare species tended to be lower in C. ciliaris plots and a

larger sample size may have yielded a statistically significant pattern.

The implications of C. ciliaris dominance for the persistence of individual species are

unknown. The abundance of individual species has been found to differ between sites sown

with C. ciliaris and unsown sites (McIvor 1998). Franks (2002) also found that the presence

of C. ciliaris affected the frequency of occurrence of some species, including Aristida spp.

and Tripogon loliiformis that declined in frequency with increasing C. ciliaris, as found here.

In study one, some species showed an uneven distribution, being found in most non-C.

ciliaris plots but in few or no C. ciliaris plots. In contrast, in study two the occurrence of

individual species appeared to be evenly distributed between C. ciliaris and B. ewartiana-

dominated plots at the 1 m2 scale (although there were not enough data to be analysed

statistically). From these studies it is unclear whether the persistence of any native species is

threatened by C. ciliaris invasion. Species richness is often an inadequate indicator of

changes in plant species assemblages (Critchley and Poulton 1998) and may therefore be a

somewhat insensitive measure of the consequences of C. ciliaris invasion. Measures of

species abundance (numbers of individuals of species present) rather than richness may be

more informative and an understanding of the effects of C. ciliaris on species abundance

may enable better prediction of the implications of C. ciliaris invasion on the long-term

persistence of individual species.

Although exotic pasture species are the basis of all virtually sown pastures in Australia and

have provided considerable economic benefits (McIvor and McIntrye 1997), the very

characteristics that make them successful pasture introductions also make them potential

weeds. Cenchrus ciliaris is considered one of the most destructive introduced grasses in

58

Australia (Low 1997) and the findings of these two studies are consistent with the view that

invasion by C. ciliaris poses a threat to biodiversity. However, further work is required to

document the ecological effects of this species. Although C. ciliaris has been found to be

associated with reduced species richness at the scales investigated, it is not clear whether this

pattern holds at larger scales. Are native species able to persist, perhaps at lower abundances,

in C. ciliaris-dominated communities? Or will C. ciliaris invasion lead to considerable loss

of species over time? To resolve the issue of C. ciliaris effects on species richness the

mechanisms behind the species richness-C. ciliaris relationships must be fully understood.

59

CHAPTER 3. MANAGING CENCHRUS CILIARIS-DOMINATED

VEGETATION WITH FIRE

3.1 INTRODUCTION

Cenchrus ciliaris is recognized as one of Australia’s worst environmental weeds (State of the

Environment 1996). This vigorous, prolific seeding, drought and grazing-tolerant grass has

been widely promoted as a valuable pasture species and now dominates the herbaceous layer

of large areas of northern and inland Australia (Griffin 1993;Woinarski 2001). It is believed

to be significantly modifying ecosystems via effects on local fire regimes (Humphries et al.

1991; Latz 1991; Low 1997; Butler and Fairfax 2003) and has been associated with reduced

native species richness (Best 1998; Fairfax and Fensham 2000; QPWS 2001; Franks 2002).

Given its perceived negative impacts, its extensive distribution in arid and semi-arid habitats

and its potential to spread, strategies for containing or preventing the spread of C. ciliaris

into non-target areas are now being sought (for example in the Alice Springs Desert Park,

see Pitts and Albrecht 2000).

Fire is one of the few tools available for manipulating plant community composition in

extensively managed vegetation. It has been used to manage tree-grass dynamics, maintain

pasture condition, increase the availability of nutritious herbage to cattle, manage grazing

distribution and reduce the hazards of wild fire (Tothill 1971; Dyer 2000). Fire also has a

role in the control of invasive species and has been investigated as a tool for controlling

invasive woody (Bebawi et al. 2000; Campbell and Setter 2002) and herbaceous (Parsons

and Stohlgen 1989; DiTomaso et al. 1999) species.

Fire has been suggested as a management tool to maintain or restore Hawaiian grasslands

invaded by African grasses such as C. ciliaris (Daehler and Carino 1998). However, there is

little information available regarding the use of fire for managing C. ciliaris in Australia. In

the past, interest in C. ciliaris-fire interactions focused on the development of ‘improved’

pastures and, generally, fire was not seen as particularly useful for developing or maintaining

C. ciliaris pastures (McIvor and Gardener 1981; ’t Mannetje et al. 1983, Pressland and

Graham 1989). In contrast, in the more recent literature discussing C. ciliaris as an invasive

species, fire is reported to favour C. ciliaris (Humphries et al. 1991; Lazarides et al. 1997;

Butler and Fairfax 2003). Many invasive grasses are associated with changes in fire regimes

(D’Antonio and Vitousek 1992; D’Antonio 2000; Wilson and Mudita 2000; Rossiter 2003)

and the fire-promoting properties of C. ciliaris are highlighted in the invasive plant literature

(Humphries et al. 1991; Low 1997).

60

The use of fire to control fire-adapted species such as C. ciliaris (refer section 3.2.4) may

initially appear inappropriate. However, even fire-adapted species are affected adversely by

fire at particular times and conditions (Vogl 1974). Manipulation of the fire regime may

offer opportunities to alter competitive interactions and disadvantage C. ciliaris relative to

the native species. Fires can be implemented at different frequencies, intensities and in

different seasons (Bond and van Wilgen 1996) and alteration of the fire regime can result in

significant changes in vegetation responses. In addition, species growing together may

respond very differently to the same fire (Daubenmire 1968) due to differences in

morphology, phenology and regeneration and/or recruitment strategies. Consequently, the

strategic use of fire may promote some species over others.

A key factor influencing the effects of fire is the season of burning. Both fire and vegetation

characteristics vary seasonally and the effects of fire may be dramatically different,

depending on its timing. Fire characteristics such as intensity and patchiness vary seasonally

with changes in fuel and weather conditions. Late dry season fires are generally more intense

and extensive than early dry season fires (Gill et al. 1996; Williams et al. 1997) since fuel

moisture levels are lower and weather conditions are more favourable for high intensity fires

at this time (Gill et al. 1996). Vegetation characteristics, such as plant size, moisture content

and phenological condition, also vary over time and plants will be more or less susceptible to

fire, depending on their life cycle and condition in relation to the timing of fire. Changes in

fire characteristics and changes in the relative susceptibility of species to fire mean that the

season of burning may have a dramatic effect on the composition and structure of grasslands

(Collins and Gibson 1990).

The use of fire to manipulate vegetation composition requires an understanding of fire

characteristics and the ecological consequences of particular fire regimes (Hodgkinson et al.

1984). Fire has the potential to change vegetation composition by affecting recruitment and

establishment patterns. As well as directly killing seedlings, fire affects recruitment and

establishment patterns via its effects on the availability of (1) sites suitable for plant

establishment and (2) propagules to colonize these sites. In perennial grasslands, the resident

plants sequester resources and major disturbances such as fire may be required to eliminate

or reduce competition and free resources for the establishment of new individuals (Cheplick

1998). In addition, fire may affect site availability by altering micro-site conditions. After

burning, light levels at the soil surface are increased and soil chemistry, biological activity

and soil and air temperatures may be altered (Bond and van Wilgen 1996). Fire affects

propagule availability by killing seed held in seed heads and in the soil seed bank. It may

also affect propagule availability via effects on flowering and seed germinability. The effects

61

of fire on establishment sites and propagules depend, in part, on the timing of fire. Burning

in different seasons may result in significant differences in site and propagule availability.

For example, late dry season fires may open up more sites for plant establishment than early

dry season fires if the higher intensity of these fires results in greater plant mortality and/or

more extensive removal of cover. Early dry season fires may destroy more seed than late dry

season fires if seed is still be held in seed heads or is on the soil surface at the time of

burning.

Can fire be used to reduce the abundance of C. ciliaris and promote the recruitment of native

species? This chapter describes a series of studies investigating the effects of season of

burning on two C. ciliaris-dominated communities in northern Queensland. The aim of these

studies was to investigate the effects of fire on establishment sites, propagules and on

herbaceous community composition to help evaluate fire as a tool to manipulate the

composition of C. ciliaris-dominated grassland. Experimental plots were established in two

communities, a Eucalyptus savanna (Dalrymple) and an Astrebla (Mitchell grass) grassland

(Moorrinya), and three burning treatments were imposed: early dry season burn (June), late

dry season burn (November) and no burn (control). Early and late dry season fires in these

communities were expected to generate significantly different fire intensities.

Three studies were conducted to investigate effects of season of burning on establishment

site availability:

(1). Can fire kill Cenchrus ciliaris (section 3.4)?

In this study the effects of season of burning on perennial plant persistence were

investigated by assessing the survival of tagged C. ciliaris and Astrebla plants after

the implementation of burning treatments

(2). Does fire increase plant nutrient availability in Cenchrus ciliaris dominated grassland

(section 3.5)?

A bioassay technique was used to investigate season of burning effects on plant

nutrient availability.

(3). Litter - a help or hindrance to seedling emergence (section 3.6)?

The effects of soil surface cover on C. ciliaris and Heteropogon contortus (a native

perennial grass) seedling emergence were investigated to help predict the effects of

litter removal by fire on seedling recruitment patterns.

62

Two studies were conducted to investigate the effects of season of burning on propagule

supply:

(1). Effects of season of burning on soil seed banks of Cenchrus ciliaris-dominated grassland

(section 3.7).

A germination method was used to investigate the effects of season of burning on

the soil seed banks of C. ciliaris-dominated vegetation.

(2). Does fire promote flowering in Cenchrus ciliaris (section 3.8)?

The aim of this study was to quantify the effects of season of burning on flowering

in C. ciliaris and Astrebla.

The effects of fire on establishment site and propagule availability may be reflected in

seedling emergence patterns and plant community composition. Difficulties in access to field

sites due to weather and resource limitations prevented a detailed investigation of seedling

emergence patterns in response to season of burning. However, two seedling studies were

conducted at the Dalrymple site (section 3.9). In the first study the pattern of forb and grass

seedling emergence was assessed early in the growing season after the late dry season fires.

In the second study grass seedling emergence was monitored over the course of the

following growing season. The aim of these studies was to determine if season of burning

altered seedling emergence patterns. The effects of season of burning on plant community

composition were investigated at both sites by recording the herbaceous species present and

their cover pre and post fire (section 3.10).

Treatment comparisons were made between C. ciliaris and Astrebla at Moorrinya because

Astrebla was co-dominant at this site. Treatment comparisons were made between C. ciliaris

and H. contortus because H. contortus was a common native at Dalrymple whereas Astrebla

was not present.

This chapter consists of 11 sections. Following this introduction (section 3.1) there is a brief

review of the use of fire as a tool for manipulating vegetation composition (section 3.2). A

description of the experimental sites and fire treatments is presented in section 3.3. The

individual studies are then reported (sections 3.4-3.10). For each study, the specific methods

used are described. However, site and fire treatment details are not given as these have been

described in section 3.3 and are common for all studies. In the final section of this chapter

(the general discussion, section 3.11), the implications of the findings for the development of

strategies using fire as a tool to manipulate C. ciliaris-dominated grasslands are discussed.

63

3.2. LITERATURE REVIEW:

THE ROLE OF FIRE AS A VEGETATION MANAGEMENT TOOL

3.2.1. INTRODUCTION

Fire is of global importance in natural and managed ecosystems, both as a destructive force

and as a management tool (Raison 1979). As a management tool, it is one of the few

available for managing vegetation on a broad scale and has been used by humans for

thousands of years to manipulate communities of plants and animals in many different parts

of the world (Christensen and Burrows 1986). In pre-modern times, fire was used in

Australia and elsewhere for a wide variety of purposes such as to clear undergrowth, hunt

game, engage in warfare and signal presence (Wilson and Mudita 2000). More recently, it

has been used to aid the establishment of improved pastures, maintain pasture condition,

increase the availability of nutritious herbage to cattle, manipulate grazing

patterns/distribution, control diseases and pests such as ticks, and reduce the hazard of

uncontrolled wildfires (West 1965; Tothill 1971; Dyer et al. 1997; Dyer 2000). Fire has also

been used to manipulate tree-grass dynamics as well as herbaceous community composition

to reduce the abundance of undesirable, invasive species (e.g. Whisenant 1990b; Bebawi et

al. 2000). The use of fire in northern Australian communities is reviewed elsewhere (Tothill

1971; Leigh and Noble 1981; Hodgkinson et al. 1984; Dyer et al. 1997; Grice and Slatter

1997)

“Fire is a powerful and rapidly acting modifier of the environment” (Raison 1979). It directly

affects plant growth, survival and reproduction and is one of the few natural disturbances

that, alone or in combination with other forces, regularly kills mature plants (Bond and van

Wilgen 1996). Fire affects all stages of the plant life cycle (Lunt and Morgan 2002),

impacting upon the vigour of individual plants and their regenerative and/or recruitment

capacity (Copeland et al. 2002). At a community level, it may alter composition and

structure via effects on seed banks, establishment site availability, seedling survival and

competitive interactions. Fire is an important factor structuring plant communities (Tyler

1995); the establishment sites it creates provide the potential for vegetation change (Bond

and van Wilgen 1996).

The effects of fire on vegetation are determined by complex interactions between fire

characteristics, plant characteristics, soils, grazing regimes and climate (Walker et al. 1981).

Fires vary in type and behaviour and differences in fire regime (frequency, intensity and

64

timing) can result in drastically different plant responses. Plants vary in morphology,

phenology and life strategy and, consequently, species differ in their tolerance of particular

fire regimes. In this section the characteristics of fire, its effects on the biotic and abiotic

environment, and plant responses to fire are briefly described with emphasis on

grassland/savanna systems. The role of fire in the control of invasive species is reviewed and

the relationship between C. ciliaris and fire is described.

3.2.2. FIRE: AN ECOLOGICAL PHENOMENON

Fire is a natural environmental phenomenon in many Australian ecosystems (Gill 1975) and

most of Australia’s 77x107 hectares is subject to its influence (McArthur 1972). Fire has long

been an important ecological factor in northern Australia, probably contributing to the

development of the region’s plant and animal communities even before occupation by

humans (Hodgkinson et al. 1984; Grice and Slatter 1997). The arrival of Aborigines and,

more recently, Europeans, has seen changes in fire regimes (Grice and Slatter 1997; Noble

and Grice 2002). As well as changes over time, fire regimes vary over the landscape;

different regimes occur in different vegetation types and in different management systems

(Gill et al. 1990). Today fire is widely acknowledged as an important ecological factor in

many Australian plant communities (Dyer et al. 1997) and fire management is recognized as

important for the conservation and management of northern Australian ecosystems (Gill et

al. 1990).

For fire to occur there must be a source of ignition, sufficient fuel and suitable environmental

(weather) conditions (Bond and van Wilgen 1996). Differences in vegetation, topography

and climate result in different types of fire and different fire behaviour.

3.2.2.1. Fire type

Cheney and Sullivan (1997) defined three types of fire on the basis of the orientation of the

edge of the fire with respect to the wind. Heading fires are fires where the flames are blown

towards the fuel, backing fires are those that move into the wind with flames leaning over the

burnt ground and flanking fires are those where the fire edge is parallel to the direction of the

wind. These fire types differ in the way they burn and different fire types will occur at

65

different positions in the one fire (Cheney and Sullivan 1997). Fires can also be categorized

as ground, surface or crown fires, each having different burning characteristics (Bond and

van Wilgen 1996). The most common types of fires in savanna areas are surface fires,

although crown fires do occur in extreme conditions (Trollope 1997). In northern Australia

crown fires are virtually unknown (Stocker and Mott 1981).

3.2.2.2. Fire behaviour

Fire behaviour refers to the physical attributes of individual fires (Williams and Cook 2001)

including rate of forward spread, flame zone characteristics (flame height, length and angle,

residence and smoulder time) (Cheney and Sullivan 1997), scorch height and intensity,

temperature, extent and patchiness (McArthur and Cheney 1966; Dyer et al. 1997). The

characteristics of grass fires have been described by Daubenmire (1968) and Vogl (1974)

and the characteristics of fire in northern Australia have been described by Gill et al. (1990)

and Williams et al. (2002).

Fuel and weather conditions are the major factors influencing fire behaviour. Topographic

factors are also important since slope affects the forward rate of spread of surface fires

(Trollope 1997). The amount, chemical properties, size and spatial arrangement of plant

material greatly influences how it will burn and these fuel characteristics vary considerably

between ecosystems (Bond and van Wilgen 1996). Fuel dynamics have been described for

Australian vegetation in general (Walker 1981) and for northern Australian vegetation in

particular (Williams and Cook 2001). The minimum amount of fuel required to carry a fire

depends on fuel distribution and moisture content (Walker et al. 1981; Hodgkinson et al.

1984). In savannas, grass is a particularly important fuel due to its flammability, particularly

in the dry season when the grasses either die out completely (annuals) or die back to the root

stock (perennials) (Gill et al. 1990). Weather conditions also determine how and when fire

will burn. Wind speed and atmospheric conditions directly affect the spread, intensity and

behaviour of fire. Rainfall, relative humidity and temperature affect fire behaviour indirectly

via effects on fuel moisture content (Bond and van Wilgen 1996). Weather also affects fire

by influencing the growth and accumulation of fuel. ‘Fire weather’ has been described for

the wet-dry tropics by Gill et al. (1996).

66

3.2.2.3. Fire regime

The term ‘fire regime’ describes how often fire occurs (frequency), when it occurs (season)

and how fiercely it burns (intensity) (Bond and van Wilgen 1996). Natural fire regimes are

determined by ignition probability and type (Williams and Cook 2001), by the amount,

seasonality and reliability of rainfall, and by the nature of the plant community (Dyer et al.

1997).

Generally the term ‘fire frequency’ is used to describe the time interval between successive

fires at a particular location (fire return interval), although it may also be used to refer to the

number of fires per unit area over a landscape (regional frequency) (Christensen 1985).

Potential fire frequency is controlled by fuel load and condition that, in turn, are determined

by factors such as vegetation type and rainfall (Walker et al. 1981). Fires of different

frequency may have different characteristics. For example, annual fires have been found to

have lower temperatures and shorter durations of surface heating than less frequent fires.

These effects are related to differences in fuel loads (refer Morgan 1999).

Season of burning defines the timing of fire in relation to climatic and vegetation conditions.

The timing of fire may greatly influence its behaviour. For example, fires in the early dry

season in savanna in the Northern Territory are patchy, of low intensity and tend to go out at

night. In contrast, late dry season fires are of higher intensity, remove grassy fuels over

extensive areas and persist overnight (Gill et al. 1990). The fuel and meteorological

conditions that influence seasonal changes in fire behaviour are described by Hoare (1985)

and Gill et al. (1996). The length of fire season varies between communities (Walker et al.

1981) depending on plant community type, climate and management practices. Wild fires in

Australia’s savannas generally occur from March to December (dry season fires), although

prescribed fires may be imposed in the early wet season (Williams et al. 2002).

Fire intensity (I) is the rate of energy release or rate of heat release per unit time per unit

length of fire front:

I = Hωr

where H is the heat yield of the fuel burnt (kJ/kg), ω is the amount of fuel consumed (kg/m2)

and r is the rate of spread (m/sec) (Cheney and Sullivan 1997). The heat yield is the amount

of heat released during a fire and is somewhat less than the heat of combustion because some

energy remains in partially burnt products (Cheney and Sullivan 1997) and there are heat

losses resulting from radiation, vaporization of moisture and other processes (Byram 1959).

67

Heat yields of grass fuels range from 11500-18600 kJ/kg (Cheney and Sullivan 1997).

The concept of fire intensity was developed by Byram (1959) (McArthur and Cheney 1966)

and is known as ‘fire line intensity’. ‘Fire intensity’ has also been used to refer to peak flame

temperature, maximum soil temperature and flame height (Walker et al. 1981). Another

measure of intensity, ‘Rothermal’s reaction intensity,’ has been defined as the rate of heat

released per unit area of the flaming zone (Rothermal 1972).

Fire intensity is strongly affected by fuel moisture as well as fuel load. Fuel moisture affects

the ease of ignition and the combustion rate (Trollope 1997) and heat yields are generally

lower at higher fuel moisture contents (Cheney and Sullivan 1997). In addition, fuel moisture

affects the rate of spread with levels above 40% usually preventing the spread of grassland

fires (Walker et al. 1981). The rate of spread is also determined by slope (Trollop 1997),

wind speed and relative humidity (Williams and Cook 2001). In grasslands, wind speed is

critical where grass tussocks are separated and there is little herbage between the discrete

plants (Walker et al. 1981).

Fires in grasslands are generally less intense than those in forest communities (Williams J. R.

1995). Consequently, the range of fire intensities is less for grasslands and savannas than for

woodlands (Bond and van Wilgen 1996). Most fire intensities in Australian’s northern

savannas range from 500-10,000 kWm-1 with intensities in excess of 20,000 kWm-1 being

rare (Williams et al. 2002). In contrast, fire intensities of up to 100,000 kWm-1 have been

reported for eucalypt forest fires (Gill and Knight 1991). There is a close link between fire

intensity and fire frequency at a site with intensity being inversely related to frequency due

to the availability of fuel (Whelan 1995). However, recovery rates influence this relationship.

Fire intensity also varies with season of burning (Lonsdale and Braithwaite 1991; Williams

et al. 1997), being affected by seasonal influences on weather conditions, particularly wind

speeds, temperature, humidity, fuel moisture content and fuel load and arrangement (Gill et

al. 1996; refer Dyer et al. 1997). The intensities of fires generally increase over the dry

season as daily wind patterns change, maximum temperatures increase, relative humidity

decreases and fuel moisture levels decline (Gill et al. 1996).

Although the fire regime is usually described in terms of frequency, season and intensity

(Gill 1975), other factors such as fire type (Trollope 1997) and extensiveness and patchiness

(Whelan 1995; Williams and Cook 2001) have also been listed as components of the fire

regime.

68

3.2.2.4. The effects of fire

Fire is the most ubiquitous, terrestrial disturbance after human urban and agricultural

activities (Bond and van Wilgen 1996). It affects a large range of plant communities, has an

extremely short period of occurrence at any point, may consume a large portion of the

standing vegetation and has a self-propagating tendency (Gill 1975). Most fires have both

beneficial and detrimental effects within an ecosystem (Raison 1980).

Fire affects plants directly by consuming or damaging biomass. It kills plants and may

destroy seed held in the standing vegetation or in or on the soil. In addition, the effects of

heat and smoke may affect flowering and seed germinability (Enright et al. 1997; Williams

et al. 2003a).

Fire affects plants indirectly by altering abiotic and biotic factors that influence plant growth.

Light, temperature, water and nutrient conditions are often markedly affected by fire. The

removal of living plant shoots and litter, that had previously intercepted much of the direct

solar radiation and retarded loss of heat by radiation (Daubenmire 1968), results in increased

light levels at the soil surface (D’Antonio et al. 2001). Removal of surface cover can also

cause major changes to temperatures of the soil surface and the adjacent layer of air

(Daubenmire 1968; Hulbert 1988). Post-fire soil and air temperatures are generally increased

(Tothill 1969; Hulbert 1988) and there may be greater fluctuations in temperature (refer

Christensen 1985; Auld and Bradstock 1996). Soil moisture and surface humidity are also

altered by fire (refer Christensen 1985; Whelan 1995). Moisture availability may increase,

due to reduced transpiration (Bond and van Wilgen 1996) and reduced loss via absorption by

litter (Daubenmire 1968), or decrease (refer Hulbert 1988) due to increased evaporative loss

(Daubenmire 1968) and to increased exposure to raindrops, leading to surface sealing and

reduced hydraulic conductivity (Bridge et al. 1983). Fire has also been found to reduce soil

wettability and hence water penetration into soils (Raison 1980). Nutrient availability is

generally increased by fire as nutrients are added to the soil surface as ash (Bond and van

Wilgen 1996). However, some nutrients are lost through volatilization and as particulates

through air and water movement (Kellman et al. 1985). Nutrient availability may be altered

as increased temperatures and changes in chemical conditions promote increased biological

mineralization (Christensen and Muller 1975; Stock and Lewis 1986). It has also been

suggested that fire alters the chemical environment via the creation of charred wood products

(Keeley et al. 1985) and by destroying allelochemicals (refer Bell 1999), although the

prevalence of this effect is uncertain (Christensen 1985).

69

Fire also affects plants indirectly via its effects on microbial and animal communities. Fire-

induced changes in temperature, moisture and pH affect microbial populations, altering

nutrient cycling (refer Raison 1979) and pathogen activity. Fire also transforms the habitats

of animals (Daubenmire 1968). The post fire environment may be hostile for invertebrates

and vertebrates including granivorous species such as ants and rats (Bond and van Wilgen

1996), altering levels of seed predation and herbivory (Christensen and Muller 1975; Tyler

1995). For example, seed predation has been found to be lower in burnt than in unburnt

grassland in southern Australia (Watson 1995).

3.2.2.5. Plant responses to fire

Fire affects plant growth and phenology. It influences the regenerative and recruitment

capacity of plants via effects on sprouting, flowering and seed set, seed viability and

germinability, and seedling establishment.

Although fire kills some plants, it may alter or even promote the growth of surviving

individuals. Many species are able to sprout after fire and surviving individuals may start

growing earlier and faster (Hulbert 1988). Fire has been reported to enhance growth and

tillering in grasses (Cheplick and Quinn 1988; refer Collins 1990). It may also affect the size

of plant organs and canopy structure (Daubenmire 1968). For example, burning was reported

to increase tiller number but reduce tiller weight in Astrebla spp. (Scanlan 1980) and

increase seed size in Nassella pulchra (Dyer 2002). New foliage may be held more erect on

burned than on unburned grassland (O’Connor and Powell 1963) and high intensity fires can

alter tree architecture, causing a reduction in foliage height (Bond and van Wilgen 1996;

refer Anderson et al. 1998).

Fire affects plant recruitment by altering propagule supply. It destroys seed held in seed

heads as well in the seed bank (Gardener 1980) and seed that is not consumed may be

damaged by overheating (’t Mannetje et al. 1983). However, although fire may decrease the

total amount of seed, it may increase the amount of germinable seed and fire-promoted

germination has been reported for many plant groups (Purdie and Slatyer 1976; Shea et al.

1979; Bell 1999; Williams 2000). Germination may be promoted by heat. Many hard-seeded

Australian species require a heat shock to permit germination (Bell 1999) and the importance

of heat in breaking dormancy has been illustrated in several plant families including

Fabaceae, Lamiaceae, Rhamnaceae (refer Auld and O’Connell 1991). For example, the

‘hard-seededness’ of legume species is broken by heat which allows water to enter the seed

70

and germination to begin (Bond and van Wilgen 1996). The degree of heat scarification

depends on the amount of seed present at various soil depths, the amount and depth of fine

fuel present, and the residence time of the fires (Noble and Grice 2002). Heat-promoted

germination may result from the heat generated in the fire itself and/or from raised soil

temperatures due to the removal of surface cover (Tothill 1969). Changes in other factors

such as light and the chemical micro-environment around the seed may also promote

germination (refer Bell 1999; refer Parker and Kelly 1999). For example, smoke has been

associated with increased germination (Roche et al. 1998; Read et al. 2000; Williams et al.

2003a). However, its importance in stimulating germination of semi-arid and tropical plant

species has yet to be determined on a comprehensive scale (Noble and Grice 2002).

Fire also affects propagule supply via effects on flowering, with both increases and decreases

in post-fire flowering being found (refer Daubenmire 1968). The association between

flowering and fire ranges from near obligate in some herbaceous species to weakly

facultative in sprouting shrubs (Bond and van Wilgen 1996). Fire-stimulated flowering is

very common in monocotyledons including Poaceae and Cyperaceae and increased

flowering after burning has been reported for grasses including Astrebla spp. (Scanlan 1980),

Amphicarpum purshii (Cheplick and Quinn 1988), Andropogon spp. (Hulbert 1988), and

tallgrass prairie grasses (Collins 1990). Flowering in other grassland plants, such as some

forbs, is also be promoted by fire (Lunt 1994). The stimulus to flower varies and has been

attributed to smoke or changes in temperature, light and nutrients (refer Bond and van

Wilgen 1996; refer Lunt and Morgan 2002). The effects of burning on flowering are likely to

alter seed production and input to the seed bank. Fire has been reported to increase seed

production in grasses (refer Orr et al. 1991; Scanlan 1980), although decreases in grass seed

production due to fire have also been reported (Whisenant 1990b). In some species fire

stimulates seed release. Such serotinous species include many woody species found in

Australian shrublands (Bond and van Wilgen 1996).

The effects of fire on seed banks and flowering flow on to affect seedling emergence

patterns. Burning often results in greater and/or earlier emergence (Daubenmire 1968) and

flushes of seedling emergence after fire are commonly reported (Christensen and Muller

1975; Shea et al. 1979; Peart 1984; Robertson et al. 1999). However, some studies have

found no evidence of fire-stimulated germination (Collins and Gibson 1990). For example,

in temperate Australian grasslands, mass germination of most perennial species does not

occur in response to single fires, probably because of small soil seed banks (Lunt and

Morgan 2002). Changes in seedling emergence patterns may result from the direct effects of

fire on seed germinability and/or from the environmental changes in the burnt area such as

71

decreased soil cover, reduced competition for resources and reduced herbivory (Tyler 1995).

Increased emergence after fire of some shrubs of wet sclerophyll forest in north-eastern

Australia is believed to result from the heat generated by fire (Williams 2000). In contrast,

Heteropogon contortus seedling emergence associated with burning (Shaw 1957) was

attributed to post-fire increases in soil temperature (Tothill 1969). Removal of shade by fire

was also considered important in influencing seedling emergence patterns in Themeda

triandra grasslands in south-eastern Australia (Morgan 1998). Fire results in reduced

emergence of species whose seeds are killed by fire (Peart 1984) or where fire creates micro-

site conditions unfavourable for seedlings.

Seedling establishment patterns are also altered by fire. In the post-fire environment where

resources such as light, water and nutrients increase, and seed and seedling predation decline

(Bond and van Wilgen 1996), seedling survival may be enhanced. For example, Tyler (1995)

concluded that increased seedling survival in burnt chaparral vegetation was associated with

reduced herbivory following fire. Given suitable growing conditions, seedling survival is

generally enhanced where fire results in reduced competition from established plants.

At the community level, fire can result in changes in vegetation composition and structure

(Cheal 1996). Some species are favoured by a particular fire regime, increasing in

abundance, while other species are reduced in abundance or are eliminated altogether.

Disturbances such as fire may promote coexistence by preventing competitive exclusion

(Morgan 1999; Valone and Kelt 1999; Lunt and Morgan 2002) and many studies report

increased species richness after fire (Parsons and Stohlgren 1989; DiTomaso et al. 1999;

Valone and Kelt 1999; Copeland et al. 2002; Williams et al. 2003b). However, the reverse

has also been reported: annual burning of tallgrass prairie increased native grass dominance

and depressed native and exotic species richness (Smith and Knapp 1999). Fire also affects

site productivity. The reported effects of fire on grassland biomass production vary. Vogl

(1974) stated that burning has generally been found to increase the productivity of grassland.

In contrast others report that burning generally reduces the biomass and cover of dominant

grasses and other species in the post-fire growing season (Tothill 1971; Lunt and Morgan

2002). Post-fire biomass production is strongly dependent on seasonal rainfall (Orr et al.

1991) and variations in rainfall and other factors, such as nutrient availability, are important

in determining biomass production.

Plant responses to fire are described in detail in Bond and van Wilgen (1996). The responses

of grassland plants to fire are reviewed by Daubenmire (1968) and the effects of fire on

vegetation of the wet-dry tropics of Australia have been reviewed by Gill et al. (1990).

72

3.2.3. FIRE AS A MANAGEMENT TOOL FOR INVASIVE PLANT CONTROL

Fire is used to manage vegetation on a broad scale to achieve various land management

goals. One of its main uses has been to manipulate vegetation composition to control

unwanted species and it has been used to control both invasive woody and herbaceous plants.

Fire is an important factor influencing tree-grass dynamics, with regular fires tending to act

against woody plants and favour grasses (Wilson and Mudita 2000). The absence of regular

fire has enabled woody species to increase in density in many ecosystems (Hodgkinson et al.

1984; Dyer et al. 1997) and the invasion of exotic and native woody species is now a major

problem in Australia’s rangelands (Russell-Smith et al. 2000; Noble and Grice 2002). Fire

can result in high mortality of woody plant seedlings and fire regimes can be designed to

increase or decrease the abundance of woody plants once life histories and fire-survival

strategies of individual species are understood (Hodgkinson et al. 1984). Fire has been used

to control invasive shrubs including Prosopis pallida (mesquite) (Campbell and Setter 2002),

Cryptostegia grandiflora (Bebawi et al. 2000), Cytisus scoparius (Robertson et al. 1999),

Acacia sophorae (McMahon et al. 1996) and Genista monspessulana (Alexander and

D’Antonio 2003). Although most woody species in tropical eucalypt forests and woodlands

are vulnerable to fire at the seedling stage (Dyer et al. 1997; refer Noble and Grice 2002),

many trees in tropical Australia can regenerate after fire (Gill et al. 1990). Therefore, high

fire intensities may be needed to kill mature individuals (Williams J. R. 1995; Dyer et al.

1997).

Fire may also be useful for manipulating the composition of herbaceous communities and

has been investigated as a tool to reduce the dominance of undesirable grasses such as

Elymus caput-medusa (Furbush 1953), Setaria (Davidson 1951), exotic annual grasses

including Avena, Bromus and Festuca species in California (Parsons and Stohlgen 1989),

Bromus japonicus (Whisenant 1990b) and Aristida spp. (Orr et al. 1991; Dyer et al. 1997). It

has also been used to manipulate problem forb species, for example Watsonia spp. (Groves

1991) and Centaurea solstitialis (yellow starthistle) (DiTomaso et al. 1999).

The use of fire as a management tool is not simply a matter of deciding whether to burn or

not (Bond and van Wilgen 1996). It requires an understanding of fire characteristics and the

ecological consequences of particular fire regimes (Hodgkinson et al. 1984). Fires can be

implemented at different frequencies, intensities and in different seasons and the size of

burnt areas and their location relative to unburnt areas can be varied (Bond and van Wilgen

1996). Differences in these characteristics can have a profound effect on the outcome of

73

burning. The use of fire also requires an understanding of the characteristics of vegetation

that influence its responses to burning. The effects of fire on plant communities depend on

the species present and the stage of their life cycle exposed to fire (Gill et al. 1990). Finally,

interactions between fire and other factors such as climate, soils and grazing must be

considered.

3.2.3.1. The importance of fire regime

“[All] fires are not equal” (Bond and van Wilgen 1996). Fires vary in intensity, frequency,

and seasonality and the effects of fire can differ enormously depending on the circumstances

of the particular fire (Bond and van Wilgen 1996). Variations in fire regime may result in

very different effects on vegetation dynamics and different regimes may be needed to

achieve different land management goals (Williams and Cook 2001). The frequency,

intensity, duration and seasonality of fire are considered important for assessing effects of

fire on plants (refer Auld 1986).

The frequency of burning may significantly affect vegetation composition and structure,

particularly where composition is strongly influenced by biomass accumulation (Lunt and

Morgan 2002) or where plant life cycles can be severely interrupted (Vogl 1975). For

example, fire frequency, rather than other attributes such as intensity, has the most profound

influence on the composition of temperate, lowland grasslands in southern Australia (Lunt

and Morgan 2002). This is because fire frequency, rather than intensity, has the greatest

effect on biomass accumulation and biomass accumulation is a major determinant of

composition in these grasslands (Lunt and Morgan 2002). Frequent burning generally

favours sprouting perennials over non-sprouting species, disadvantages species that rely

solely on on-site storage of seed, promotes herbaceous over woody species, and promotes

grasses over dicotyledons (Vogl 1977). Communities, such as mulga shrublands, that rely on

regeneration from seed following the mortality of adults are particularly susceptible to

changes in fire frequency. In some communities, fires in close succession can totally

eliminate particular species (Griffin and Hodgkinson 1986) and cause complete conversion

of community type (Vogl 1975). Frequent burning of heathlands has been found to reduce

both structural and floristic diversity (Cheal 1996) and in savanna woodlands, the abundance

of individual species varies with changes in fire frequency (Williams et al. 2003b). Fire

frequency may be important in altering the success of colonizing species (Baird 1977;

Laterra 1997) and reducing fire frequency may be an effective control method for some

weeds, for example the South African grass Ehrharta calycina (Groves 1991).

74

Although fire frequency has a major effect on the composition of some communities, it

appears to be less important in others, for example some northern Australian savannas (refer

Williams et al. 2002) and tallgrass prairies in Oklahoma (Engle et al. 2000). Fire frequency

(annual versus biennial fires) was found to have little effect on the composition of tropical

woodlands in the Northern Territory (Bowman et al. 1988).

The effects of fire also depend on the number of burns. For example, exotic forbs were found

to increase after a single grassland burn in California but then decrease in cover after

repeated burns (Eller 1994). Exotic grasses, also in Californian grasslands, were reported to

decrease with burning, but only after three burns, and they quickly regained their pre-

treatment dominance if burning was halted (Parsons and Stohlgren 1989).

Fire intensity is a major factor determining vegetation responses to burning. Intensity levels

influence the proportion of biomass consumed (Williams and Cook 2001), the level of plant

mortality (Griffin and Friedel 1984a; Williams J. R. 1995), germination responses (Shea et

al. 1979), vegetation structure (Williams J. R. 1995) and plant biodiversity (Braithwaite

1995). Species differ in their tolerance of particular fire intensity levels and differences in

mortality rates can result in marked changes in relative abundance (Lonsdale and Braithwaite

1991). Generally, high intensity fires favour the regeneration of hard-seeded legumes and

limit the growth of obligate-seeding, proteaceous species (Groves 1991). Plant species within

tropical eucalypt forests and woodlands are well adapted to frequent, low-intensity fires

(Dyer et al. 1997) and unusually intense fires following favourable seasonal conditions can

increase tree mortality (Lonsdale and Braithwaite 1991) and cause significant changes in

forest structure (Williams J. R. 1995). Fire intensity also affects habitat heterogeneity. Fuel

consumption tends to be complete for savanna fires greater that 2000 kWm-1 while below

this intensity some fuel remains, creating a heterogeneous mix of burnt and unburnt patches

(Williams and Cook 2001). Patchiness in vegetation created by fire is recognized as

important for promoting diversity (Hodgkinson et al. 1984) and the patchiness of litter

removal by fire can affect community structure (Facelli and Pickett 1991a).

Although fire intensity is considered a key attribute determining the effects of fire, its

relevance in determining fire effects at the plant level has been questioned. Fire intensity

may be an inadequate indicator of the effects of fire since other factors, such as flame

dimensions, temperature profiles, smoke and fine grain patchiness left by fire, may be

biologically more important (Gill et al. 1990). Raison (1979) considered that fire intensity

may not be useful for describing the heat pulse that produces the ecological effect and Noble

and Grice (2002) stated that the duration of fire, or fire residence time, is probably more

75

important than fire intensity. The length of time required to kill plant tissue is exponentially

and inversely related to temperature (Bond and van Wilgen 1996) and therefore, the length

of time fire is adjacent to a plant (residence time) is critical to plant survival (Walker et al.

1981). Much of the variation in plant responses to burning can be attributed to the varying

sensitivity of different species or tissues to heating (Bond and van Wilgen 1996). Flame

height in relation to bud position is also important. Often the height of maximum heat in

grass fires is well above the ground surface such that low placed buds and seeds escape

damage and head fires may cause less damage than backfires because the maximum

temperature is well above the ground (Daubenmire 1968). Whether plants actually respond

to ‘fire line intensity’ or to other related fire characteristics remains unclear (Noble and Grice

2002).

Another key determinant of the effects of fire is the season of burning. As noted by

Daubenmire (1968), “the time of year, or even the time of day, when the fire occurs is almost

as important at the occurrence of the fire itself. Complete reversals of the type of influence

may result simply from differences in timing”. Season of burning can affect herbaceous

community composition (Tothill 1971; Mott and Andrew 1985a) and structure (Lane and

Williams 1997). For example, summer fires in central Australia were found to reduce the

yield of palatable grasses whereas winter fires either maintained or increased them (Griffin

and Friedel 1984b). Season of burning has also been found to affect tree seedling

recruitment, tree sprouting (refer Bond and van Wilgen 1996) and tree mortality. For

example, burning mallee eucalypts in autumn causes much greater mortality than spring

burning (Noble 1997).

The timing of fire is important since both fire characteristics and the condition of plants and

their susceptibility to fire vary seasonally. The effects of season of burning may be attributed

to seasonal variation in fire intensity as a result of seasonal variations in fuel and weather

conditions (section 3.2.2.2). In Northern Territory savannas, tree mortality may be higher in

late dry season fires because they are generally more intense than early dry season fires

(Williams et al. 1997). Alternatively, season of burning effects may be independent of fire

intensity effects, being instead related to the condition of the vegetation. Species adapted to

being burnt in one season are likely to be adversely affected by a burn at some other time of

the year (Hodgkinson et al. 1984).

Whether or not plants are actively growing and the stage of life cycle at the time of fire can

have major effects on plant responses to burning. Fires that occur at the beginning or end of

the growing season, before plants become physiologically active, should have less effect on

76

recovery than fires in the middle of the growing season (Bond and van Wilgen 1996). The

end of the dry season has been reported to be the least injurious time for burning grassland in

South Africa (West 1965) and northern Australia (Mott and Andrew 1985ab) since the

grasses have seeded and become dormant by this time (Dyer 2000), Consequently, dry

season fires may have little effect on biomass production or composition (Mott and Andrew

1985a). In contrast, burning these communities in the wet season can result in changes in

species abundance (Smith 1960; refer Trollope 1997). Early wet season burning may kill

recently established seedlings (Hodgkinson et al. 1984) and eliminate annual grasses because

they are burnt before flowering and seed set can occur (refer Stocker and Mott 1981).

The effects of fire are also influenced by its timing in relation to other factors such as soil

moisture and cover. Seasonal changes in soil moisture may influence the effects of fire on

plant and seed survival since, the drier the soil, the higher its surface temperature when grass

burns. However, at the same time the low moisture content reduces downward conduction of

heat (Daubenmire 1968). Plants may respond to season of burning effects on nutrient

availability. For example, late dry season fires may have greater effects on nutrient stores

than early dry season fires (refer Williams et al. 2002). The timing of cover removal also

influences the outcome of burning. Run-off and erosion may be more severe following late-

dry season fires (Townsend and Douglas 2000). In north American tallgrass prairie, the

removal of litter by fires in spring may contribute to an increase in productivity and

flowering due to increased light and solar warming of the soil while removal of litter by

summer fires may result in desiccation and inhibition of regrowth (Copeland et al. 2002).

The timing of cover removal may also affect competitive interactions between species.

3.2.3.2. The importance of vegetation characteristics

Factors such as plant morphology, chemical composition, phenology and the spatial

arrangement of plants influence both fire characteristics and vegetation responses to fire.

Plant communities have a significant influence on fire regimes and behaviour. The amount

and flammability of the biomass produced by plants controls fire frequency, intensity and

seasonality. The spatial arrangement of plants influences how well fire may be carried by the

vegetation and morphological and chemical differences between species can affect rates of

spread (Walker et al. 1981). Fire, in turn, can affect all stages of the plant life cycle (Lunt

and Morgan 2002). Although fire may be considered indiscriminate in its removal of above

ground vegetation (Alexander and D’Antonio 2003), plant responses to fire can vary

considerably between and within vegetation types. Importantly, species growing together

77

may respond very differently to the same fire (Daubenmire 1968).

The survival of individual plants is affected by a number of factors including bark thickness,

crown architecture and the location of buds (Bond and van Wilgen 1996). Differences in

bark type afford differing degrees of fire protection (Lonsdale and Braithwaite 1991) with

bark thickness and its thermal properties being important (Whelan 1995). Differences in

growth form may also result in varied responses to fire. For example, it has been suggested

that the more loosely packed tussocks of Heteropogon contortus make it more susceptible to

heat damage from fire than Themeda triandra, whose dense, internally damp tussocks

effectively insulate the growing apices (Walker et al. 1981). Size influences the effects of

fire. In theory, the bulk of a seed or organ has a bearing on its susceptibility to heat damage:

the smaller the organ the more quickly it is brought to lethal temperature (Byram 1948).

Mortality due to fire has been found to be greater for smaller individuals (refer Bond and van

Wilgen 1996; Grice 1997). However, smaller diameter plants are sometimes less susceptible

to fire, presumably because the smaller volume of fuel results in less heat being released

(Wright and Klemmedson 1965). The position of buds at the time of burning is also

important. Plants with underground, perennating buds are likely to be less sensitive to fire

than those that tiller from above the ground (Daubenmire 1968) and plants whose stems

remain compact are likely to be less vulnerable than plants whose stems elongate slowly,

exposing the apex over a long period (Tainton 1981). Grasses are among the most fire

resilient components of plant communities (Bond and van Wilgen 1996). Although they may

form highly flammable fuels, they recover quickly from burning (D’Antonio and Vitousek

1992). Since the growing points of grasses are generally near or beneath the soil surface and

protected from all but extreme heat, fire destroys little more than the accumulated growth,

most of which is dry and dead when the plant is dormant. Most woody plant species in

tropical eucalypt forests and woodlands are able to sprout from protected buds in

lignotubers, lateral roots and stem bases at or below ground level and from protected aerial

epicormic buds (Dyer et al. 1997).

Plant phenology is also important in determining the effects of fire (Hodgkinson et al. 1984;

Williams et al. 2002). Plant moisture content and the location of carbohydrate stores (roots

or shoots) significantly influences plant responses to fire. Plants with higher moisture

contents are killed at lower temperatures (Bond and van Wilgen 1996). Therefore, actively

growing individuals may be highly susceptible to fire while dormant individuals are

relatively unaffected (Daubenmire 1968; Bond and van Wilgen 1996). The strong seasonal

growth patterns of many grassland species have significant implications for burning in

different seasons (Lunt and Morgan 2002). As noted above, burning when grasses are

78

dormant may have little effect on plants. At this time, carbohydrate levels are at a maximum

in root storage organs and the regenerative buds of many species are tightly held in tussock

bases, effectively insulated from direct fire damage (Lunt and Morgan 2002). Perennial

grasses will be most affected by fire when they are actively growing (Bond and van Wilgen

1996) and the major part of the food reserves have been withdrawn from the underground

organs (Aldous 1934).

Differences in the phenology of resident plant groups may result in changes in community

composition with burning since fire may damage or promote certain species. For example, in

tallgrass prairies in Wisconsin, species flowering before mid July are favoured by July burns

while late-flowering species are favoured by March (dormant season) fires (Howe 1994).

Burning annual grassland in northern Australia at the start of the wet season after the

majority of Sorghum spp. seeds have germinated but prior to flowering and seed set can

reduce sorghum abundance (Stocker and Sturtz 1966). Species also differ in the time taken to

recover reproductive capacity after fire (Hodgkinson et al. 1984) and such differences can

influence vegetation composition since species that are quick to recover gain a competitive

advantage over slower species (Daubenmire 1968).

Factors such as fire-promoted seed germination, protected seed embryos (hard seededness),

heavy seed production, early reproductive maturity, seed burial and seed longevity improve

the recruitment ability of species after fire (Walker et al. 1981). The position of seeds at the

time of burning is critical (Daubenmire 1968; Cheplick and Quinn 1988). The high

temperatures generated by fire may not penetrate deeply into the soil (Cheplick and Quinn

1988) and so, while seed on the soil surface may be killed by heat, seed buried below the

surface may be protected. Differences in seed size and shape are important in determining

the degree of burial on a particular soil surface (Grubb 1977) and consequently differences in

seed morphology will influence the seed’s vulnerability to fire (Walker et al. 1981; Peart

1984). In monsoon and tallgrass communities in northern Australia, perennial species such

as Themeda triandra and Heteropogon contortus, and annual Sorghum spp. have seeds with

twisting hydroscopic awns that help bury them into the soil thus protecting them from fire

(Dyer et al. 1997). The seeds of conifers and some other tree species are protected from the

heat of fire by woody fruits (Whelan 1995).

Plant species vary in their susceptibility to fire and in their vegetative and reproductive

responses to fire. Researchers have attempted to classify plants into fire life-history

categories based on plant survival and reproductive responses to fire (Gill 1981).

Characteristics of grassland plants that affect their reactions to fire have been described by

79

Daubenmire (1968) and traits that adapt plants to fire are reviewed for the Australian biota

by Gill (1975, 1981).

3.2.3.3. The importance of other factors

Both fire and plant characteristics are influenced by factors such as climate, topography,

soils and grazing. These factors influence fire regimes via their effects on fuel loads and

composition. They are also critical in determining post-fire vegetation responses. In fact,

interactions between fire, rainfall and grazing and/or browsing are often far more important

than the effects of fire alone (Noble and Grice 2002)

Climate has a major effect on fire-vegetation interactions. The frequency and intensity of

fires is determined by the build up and condition of fuel. Climate influences fuel loads and

type via its influence on species composition, biomass production and plant phenology.

Weather conditions directly influence fire behaviour via effects on humidity and wind speed

and post-fire weather conditions play a major role in determining vegetation responses after

fire (Mott and Andrew 1985a). Rainfall in particular influences recruitment success (Bond

and van Wilgen 1996) and post-fire productivity (Orr et al. 1991). For example, increased

growth in Astrebla grassland after fire occurred only at sites that received more than 500 mm

of rain during the growing season; below this, no change or a reduction in growth occurred

(Scanlan 1980). In some communities, the rainfall regime following fire is the dominant

influence on post-fire development (Noble 1989).

Grazing patterns may alter local fire regimes by affecting plant biomass and species

composition and changes in grazing regime are accompanied by changes in fire regime

(Christensen and Burrows 1986). The presence of livestock and/or concentrations of native

herbivores before and particularly after burning can completely alter the responses of

vegetation to fire (refer Vogl 1974). Post-fire grazing by domestic, feral and native animals

is a critical factor in determining the effects of fire on herbaceous species (Walker et al.

1981).

In addition to grazing, management practices such as the development of improved pastures

with high yielding species such as C. ciliaris can also alter fire regimes (Walker et al. 1981).

Soil type is an important determinant of vegetation composition and productivity and plays a

role in determining fuel quality and quantity. Soil type is also likely to be important in

80

determining the effects of fire on soil nutrients (Raison 1979) as well as post-fire erosion

processes (West 1965; McIvor et al. 1995). Topography also affects fire-vegetation

interactions. Slope affects fire behaviour directly (refer page 66) while differences in

nutrient, water and propagule availability between run on and run off zones influence fuel

accumulation and vegetation responses to burning.

3.2.3.4. The efficacy of fire as a tool to manage vegetation composition

Fire is an effective tool for manipulating both woody and herbaceous vegetation. It is the

cheapest tool available (Christensen and Burrows 1986) and has the advantage that it can be

easily used over large areas (Grice et al. 2000). However, it must be tailored to the objectives

and circumstances. With regard to managing invasive plants, the efficacy of fire as a

management tool depends on the invasive species and the plant community involved.

Many species are actually promoted by particular fire regimes (Grice et al. 2000; Wilson and

Mudita 2000) and generally fire promotes, rather than limits, invasive plants (see review by

D’Antonio 2000). Disturbances such as fire enhance the success of invasive species by

altering resource availability and by altering biotic interactions and community structure and

productivity (refer Smith and Knapp 1999). The invasion of fire-tolerant, exotic plants often

results in positive feedback loops between fire and weed invasion: as the weed becomes

more abundant, fire frequency and intensity increases, native species decline, the exotic

species increases in abundance, further altering the fire regime (D’Antonio and Vitousek

1992; Wilson and Mudita 2000). This is believed to be the case for various perennial,

tussock-forming grasses of mainly African origin that have been introduced into northern

Australia, including C. ciliaris (Wilson and Mudita 2000; Butler and Faifax 2003),

Andropogon gayanus (Rossiter et al. 2003) and Pennisetum polystachyon (Gill et al. 1990).

The exotic-promoting effects of fire may occur in ecosystems where fire is naturally

uncommon. For example, fire has been reported to promote exotic species in Hawaiin

woodlands. The native species in these woodlands, which have not been previously exposed

to fire, respond poorly to burning while the exotic grasses respond well (D’Antonio et al.

2001). In contrast, fire may be ineffective in promoting exotics in ecosystems where fire has

a long evolutionary history, for example Californian chaparral and mixed conifer forest,

Australian monsoonal forests and African savanna. Nevertheless, fire has been found to

promote exotic species in systems where fire has been important as a long-term ecological

force (D’Antonio 2000).

81

Hitchmough et al. (1994) concluded from the literature that “fire has often been seen as the

most appropriate means by which to manage grass weeds”. However, although fire is

available as a management tool, it is often inappropriate in practice, being ineffective or even

promoting invasive species. It may not be useful for promoting native species at the expense

of exotics where seed banks are dominated by exotic species (Lunt 1990; D’Antonio et al.

2001). D’Antonio (2000) concluded from her review of fire-plant invasion studies that fire

was often not effective in controlling invasive species and that, in some cases, it resulted in

an increase in the abundance of other non-target exotics (e.g. Parsons and Stohlgren 1989).

The value of fire as a management tool depends on the responsiveness of the plant

community to fire. Fire may be highly effective for manipulating community composition

where resident plant groups differ significantly in fire tolerance and/or phenology (Howe

1994). Season of burning may be particularly effective in changing vegetation composition

in such communities since fire damages plants at different stages in their development,

resulting in altered competitive interactions (Copeland et al. 2002). Fire may also be

effective for manipulating vegetation in communities where fire normally plays a major role

in plant dynamics. Dry season fires may have significant effects on communities containing

fire-triggered germinators (Williams et al. 2003b) but be less effective in communities where

other factors, such as edaphic conditions (Bowman et al. 1988) or climate, have a large

influence on species composition. For example, dry season fires may have little effect on

communities in which seeds of the resident species are prompted to germinate at the start of

the wet season regardless of fire (Williams et al. 2003b).

In savanna systems, the primary determinants of composition and structure are variations in

moisture and soil nutrients whereas disturbances such as fire are of secondary importance

(refer Williams et al. 2002). In these systems the usefulness of fire alone as a tool for

manipulating vegetation composition may be limited since other environmental factors are

key determinants of vegetation dynamics.

However, fire is often useful in combination with other tools and the use of fire as part of an

integrated approach to weed control, or ‘ecological control’ as described by Groves (1991),

has been advocated (Vitelli 2000; Noble and Grice 2002). Although individual weed control

methods, such as mechanical control, herbicides, burning and biological control, have their

advantages and disadvantages, no single method is usually sufficient for effective weed

management. A combination of control options implemented at appropriate times will be

more effective (Vitelli 2000). Thus burning followed by the application of herbicides has

been investigated as a control strategy for woody species in semi arid (Noble et al. 2001) and

82

temperate (Robertson et al. 1999) woodlands. In relation to invasive grasses, the chances of

developing biological control agents are slim (Wapshere 1990) and the scope for using

chemical controls limited (Grice pers. comm.). However, the use of grazing combined with

fire has been found to be effective. Fire and grazing management may interact to cause large

shifts in composition and has been used to reduce the abundance of undesirable Aristida spp.

and promote H. contortus in pastures in southern Queensland (Orr and Paton 1977). Fire and

grazing has also been used in Californian grasslands to control unwanted non-native species

(refer Dyer 2002).

As well as recognising the advantages of an integrated approach to weed management, it is

important to note that single applications of any method are unlikely to achieve control and

follow-up action is inevitably necessary (Grice et al. 2000). Regular burning may be needed

to maintain control (Parsons and Stohlegren 1989).

Determining the efficacy of fire for manipulating vegetation composition is difficult.

Climatic effects may complicate the interpretation of burning treatments (Norman 1969) and

both weather and grazing factors can make it difficult to predict vegetation changes resulting

from fire (Walker et al. 1981). On this point, Tothill (1971) stated that the “Often quoted

usefulness of fire for vegetation management is largely or wholly due to some other factor

such as grazing”. Plant community responses to fire are influenced by the many complex

interactions between plant species, previous fire regimes, fire intensity, rainfall before and

after the fire, soil type and prior or subsequent animal grazing (Walker et al. 1981) and

combinations of factors are most likely to control responses to burning (Hulbert 1988).

Recent literature on fire and exotic species suggests that fire generally tends to promote

rather than discourage invasive species (D’Antonio 2000). Nevertheless, fire may be a useful

tool for decreasing the prominence of species that may otherwise dominate the community

(Stuwe and Parsons 1977; Whisenant 1990b) and there is evidence that many introduced

species can be controlled by fire (D’Antonio 2000). However, if fire is to be used for

invasive plant control it must be carefully applied and plant and fire factors must be critically

evaluated (D’Antonio 2000). The effectiveness of fire in controlling invasive plants depends

on physiological properties of both the native community and the invading organism as well

as the fire regime itself (Christensen and Burrows 1986). The invasive species must be

susceptible to a particular fire regime and the community to be restored must be resilient to

that fire regime (Grice, unpublished). Caution is required since fire may have undesirable

side effects such as reducing resource availability (Smith and Knapp 1999) and promoting

soil erosion (Shaw 1957). In addition, there may also be costs associated with loss of grazing

83

capacity (Hodgkinson et al. 1984). In pastoral lands, the benefits from burning, in terms of

livestock production, may not be sufficient to justify the loss of forage from burning to

control small infestations of weeds (Vitelli 2000). Also, if fire is to be used as a management

tool, the environmental benefits, for example reduction of one invader, must be weighed

against the costs, for example, increases in another invader or decline in native species

(D’Antonio 2000).

3.2.3.5. Community responses to fire: creating gaps and filling them

Fire is a major factor structuring plant communities (refer Tyler 1995). It may alter

community composition, richness, and diversity as well as competitive interactions,

succession and patch structure (Collins and Gibson 1990) and its role in changing the

relative abundance of plants is broadly recognized (refer D’Antonio et al. 2001). Can fire be

used to reduce the abundance of C. ciliaris?

The aims of management strategies to control invasive plants are to prevent or contain

invasion, and where invasion has occurred, to eliminate or reduce the abundance of the

invasive species while promoting, or at least not disadvantaging, desirable species. To assess

the usefulness of fire as a tool for altering the composition of invaded grassland we need to

understand how fire affects (1) the persistence of the invasive species, (2) the availability of

sites for plant establishment, via effects on plant competition and other factors that influence

resource availability and (3) recruitment processes, via effects on seed availability and

seedling establishment patterns. It is also important to understand how the outcomes of fire

are influenced by other factors such as climate and grazing.

The role of establishment sites, or ‘ecological gaps’, in plant community dynamics has long

been recognized (Harper 1977; Cook 1979; Cook et al. 1993a). Establishment sites are

defined by the growth requirements of individual plants since resource availability is a

function of both the resources present and the ability of a plant to compete for those

resources. This is an important point; individual sites may be suitable for some species but

not for others. Establishment sites may be vacant space where a plant has died, freeing up

resources which are then available for new individuals. Establishment sites may also exist

where plants are present, but the available resources limit their growth, and new species are

able to colonize due to lower requirements for the limiting resources or an alternative

strategy for overcoming the resource limitation (Cook et al. 1993a).

84

Establishment events are rare in many perennial grasslands (Lunt and Morgan 2002;

Lauenroth and Aguilera 1998) where the persistence of the resident perennial plants

precludes the establishment of new individuals. In communities where the longevity of

individual plants is significant, habitat disturbance plays an important role in creating

openings for seed germination and seedling establishment (refer Cheplick 1998). Fire alters

the availability of establishment sites directly, by increasing or decreasing nutrient levels,

and indirectly, by removing or reducing competition from resident plants, freeing up

resources for new individuals (Cook 1984; Cheplick and Quinn 1988). Fire may also affect

the availability of establishment sites via effects on factors such as herbivory, allelopathic

chemicals and litter. For example, litter removal by fire may have significant consequences

for the persistence of some species (Whisenant 1990b). The effects of fire on establishment

site availability may result in major shifts in species composition.

The establishment of new individuals also requires a source of propagules. Fire may cause

increases or decreases in seed input via its effects on flowering. It may reduce the amount of

viable seed held in seed heads or in the soil seed bank while its effects on germinable seed

availability may be positive or negative, depending on the species involved and fire

characteristics.

Developing fire management strategies requires an understanding of the mechanisms by

which fire affects plants (Whisenant 1990b). While the fire-promoting properties of C.

ciliaris are often highlighted in the invasive plant literature (Humphries et al. 1991; Low

1997; Franks et al. 2000; Pitts and Albrecht 2000), there is little published information on

how fire affects the dynamics of C. ciliaris-dominated communities.

3.2.4. CENCHRUS CILIARIS AND FIRE

There is a divergence of opinion regarding C. ciliaris-fire interactions in Australia depending

on the context in which it is discussed. In a pastoral context, fire has not been seen as

particularly useful for managing C. ciliaris (McIvor and Gardener 1981; ’t Mannetje et al.

1983; Pressland and Graham 1989). In contrast, in literature discussing C. ciliaris as an

invasive species, fire is believed to promote C. ciliaris (Lazarides et al. 1997; Franks et al.

2000; Butler and Fairfax 2003).

85

Fire has not been considered particularly useful for C. ciliaris pasture establishment (McIvor

and Gardener 1981, 1985) or maintenance. With regard to using fire to rejuvenate C. ciliaris

pastures, Pressland and Graham (1989) reported that annual burning failed to improve grass

yield over a three year period. ’t Mannetje et al. (1983) concluded that, although fire caused

an initial reduction in green material, it had no lasting effects on the pastures. However, in

the invasive plant literature, C. ciliaris is considered to be favoured by fire and to promote it

(Humphries et al. 1991; Low 1997; Franks et al. 2000; Pitts and Albrecht 2000). Lazarides et

al. (1997) stated that C. ciliaris was “encouraged” by fire and Butler and Fairfax (2003)

found that burning increased C. ciliaris cover. The vegetative growth of C. ciliaris has been

reported to respond vigorously after fire and seed production has been reported to be prolific,

particularly if fire is followed by rain (L. Baker, pers. comm. cited in Humphries et al.

1991). In turn, domination of vegetation by C. ciliaris is reported to lead to hotter, late

season fires and increased incidence of fire (Humphries et al. 1991; Latz 1991; Butler and

Faifax 2003). These changes to fire regimes are attributed to the higher fuel loads and later

curing of C. ciliaris biomass compared with that of native species (Humphries et al. 1991) in

combination with rapid regrowth after fire (Pitt and Albrecht 2000). Cenchrus ciliaris has

been reported to produce two to three times as much biomass as native species in central

Australia (Latz 1991). In Eucalyptus populnea lands, clearing of native vegetation and

establishment of C. ciliaris has been reported to lead to increases in fuel loads of up to 20

times that found in similar intact, native vegetation (Walker et al. 1981). From their study of

C. ciliaris-dominated vegetation in Northern Territory, Pitts and Albrecht (2000) considered

fire unsuitable for reducing C. ciliaris abundance since it recovered rapidly after fire while

the germination of native species was negatively affected.

Cenchrus ciliaris is a fire-adapted species (Butler and Fairfax 2003). Many exotic invasive

species, particularly the invasive African grasses, have co-evolved with fire and some of

these species are referred to as ‘fire weeds’ since they produce much greater flammable

biomass than the native species they replace, promoting positive feedback loops between fire

and weed invasion (Wilson and Mudita 2000). However, this does not preclude fire as a

useful tool to manage these species. Although many invasive species are promoted by fire,

there is evidence that others can be controlled by fire (D’Antonio 2000). Even fire-adapted

species are adversely affected by fire at particular times and under certain conditions (Vogl

1974) and the strategic use of a controlled fire regime has been suggested as a way to

maintain or restore areas in Hawaii invaded by C. ciliaris and other invasive African grasses

(Daehler and Carino 1998). In tropical Australia, the native perennial grasses have evolved

with fire and survive, although species differ in their responses to burning (McIvor and Orr

1991). The strategic use of fire may alter the competitive interactions between C. ciliaris and

86

the resident native species.

The role of fire in either reducing or enhancing biological invasions in native communities

depends on the properties of the native community, the invading organism and the fire

regime (Christensen and Burrows 1986) and research is required to determine how fire

influences competitive interactions between native and exotic species (Daehler and Carino

1998). How might fire work to influence C. ciliaris abundance relative to other species? Can

fire be used to kill established C. ciliaris plants? How does fire affect C. ciliaris seed

availability and seedling recruitment compared with other species? Little is known about

how different fire regimes affect C. ciliaris compared with native species. In the following

sections a series of studies investigating the effects of early dry season and late dry season

burns on C. ciliaris-dominated grasslands in two vegetation communities are described.

87

3.3. SITE AND TREATMENT DESCRIPTIONS

3.3.1. STUDY SITES

Experimental plots were established in two vegetation communities in north Queensland: a

eucalypt savanna at Dalrymple National Park and a perennial grassland at Moorrinya

National Park.

3.3.1.1. Dalrymple National Park

Dalrymple National Park (190 48’30’’ S 1460 15’30’’ E) is approximately 46 km north-west

of Charters Towers (QDE 1998a), in the Dalrymple Shire, north Queensland (Figure 3.1).

Covering 1640 ha, the park conserves part of the upper Burdekin River catchment and lava

flows of the Great Basalt Wall in the Einasleigh Uplands biogeographic region (QDE

1998a). The Burdekin River, the longest river on Queensland’s east coast, flows through the

park. The geology of the area is variable and includes lava flows, limestone, sandstone and

sandy deposits. Six vegetation communities containing 198 plant species have been

identified. The vegetation is dominated by tall eucalypts (Eucalyptus tessellaris and E.

tereticornis), she-oak (Casuarina spp.) woodlands and dry vine thicket. The herbaceous

layer is dominated by exotic perennial grasses including C. ciliaris and Panicum maximum.

The diverse vegetation and reliable water provide significant habitat for animals including

183 birds, 47 reptiles and 36 mammals (QDE 1998a). The park lies within the wet-dry

tropics and has a warm, subhumid climate (Rogers et al. 1999). Average annual rainfall in

the Dalrymple Shire ranges between 500 and 700 mm, 80% of which falls between

December and April, although rainfall is extremely variable from year to year (Quirk et al.

1997).

The park was ‘dedicated’ in 1990 to conserve a diverse range of plant and animal

communities of the Burdekin River catchment (QDE 1998a). The area also has cultural

heritage values. Remains of the old Dalrymple township, the first surveyed inland settlement

in northern Australia, are surrounded by the park but are privately owned. The park was

previously a cattle station and was grazed by stock prior to August 1992 (QDE 1998a). The

extensive grazing regime maintained low to moderate fuel levels which largely precluded a

systematic prescribed burning regime and reduced the incidence of wild fires. The only

major wildfire in recent history occurred in October 1983 when 50% of the northern section

88

of the park was burnt (Paul Williams pers. comm.).

Figure 3.1. Location of Dalrymple and Moorrinya National Parks. Grey lines indicate roads.

Townsville

Charters Towers

Torrens Creek

Dalrymple National Park

Moorrinya National Park

0 50 10025 Kilometres

The site for this study was located on the western side of the Burdekin River near the site of

the old Dalrymple township. This area is a flat to gently undulating alluvial plain bordered

by river levees. The soil is an orthic Tenosol (Isbell 1996). It is a dark yellowish-brown, fine

sand grading to a brown loam. The vegetation is open eucalypt woodland with an upper

stratum dominated by Eucalyptus tessellaris and E. tereticornis. The herbaceous layer is

dominated by C. ciliaris with some native grasses present including Bothriochloa spp.,

Chrysopogon fallax, Enneopogon spp., Eragrostis spp. and Heteropogon contortus.

89

3.3.1.2. Moorrinya National Park

Moorrinya National Park (21030’ S, 145000’ E) is 85 km south of Torrens Creek, north

Queensland (Figure 3.1). Covering 32,607 ha, it comprises flat plains crossed by shallow,

intermittent watercourses. The geology is mostly Quaternary silt, clay, sand and rubble with

rises of Tertiary conglomerate (QDE 1998b). The vegetation includes grassy plains and open

eucalypt, acacia and melaleuca woodlands. Over 300 plant species have been identified

(QDE 1998b). A diverse fauna has been found in the park including 165 species of birds, 40

reptiles, 17 mammals, nine frogs and seven fish (QDE 1998b). The climate of the area is

semi-arid with a highly variable, summer dominant rainfall (DNR 1998). The mean annual

rainfall at Hughenden is 491 mm (Bureau of Meteorology, unpublished).

The park was ‘dedicated’ in 1993 and conserves a diverse range of plant and animal

communities (12 of the 15 regional ecosystems found in the Prairie-Torrens Creek alluvial

Province of the Desert Uplands biogeographic region) (QDE 1998b). Prior to 1993 Shirley

Station, as the park was formerly known, was grazed by sheep and cattle. The fire history of

the park is not well documented although fire is considered to be the major influence on

vegetation patterns (QDE 1998b).

The site used for this study is open grassland north of Bells Outstation near Alice’s creek.

The area, which was cleared and sown with C. ciliaris in the 1960s (McCallum 1998), is

dominated by C. ciliaris and Astrebla spp. Other grasses present include Dichanthium and

Iseilema species. It is likely that prior to clearing the site was a mixture of gidgee (Acacia

cambagei) and native grasses including Astrebla spp. The site is treeless, although there are

scattered shrubs (Acacia farnesiana, A. cambagei and A. shirleyi). The landform is a gently

undulating plain of grey cracking clay (grey Vertisol, Isbell 1996) with a gilgai micro-relief.

3.3.2. EXPERIMENTAL DESIGN AND TREATMENTS

Twelve plots (20 m by 20 m) were pegged out at each site in February 1999 at Dalrymple

and April 1999 at Moorrinya. At Dalrymple, nine plots were located on a level site and three

plots were located about 200 m away, on the side of a levee bank. These three plots were

positioned away from the main experimental areas to avoid heavily timbered areas, although

trees could not be completely avoided and were present in some plots. At Moorrinya, 12

plots were set up in a three by four grid within an area (approximately 100 m by 150 m) of

grassland. Although treeless, some plots contained shrubs such as A. cambagei, A.

90

farnesiana and A. shirleyi. Fire breaks were slashed around the perimeters of all plots prior

to burning.

The treatments imposed were: early dry season burn, late dry season burn and no burn

(control). These treatments were chosen to achieve significantly different fire intensities.

While an early wet season burn treatment was also of interest, wet season fires are

logistically more difficult to achieve and resources were not available to implement them.

Early dry season fires were conducted on 10th and 15th of June 1999 at Moorrinya and

Dalrymple respectively. Late dry season fires were conducted on the 12th and 17th of

November 1999 at Dalrymple and Moorrinya respectively. The wet season in this region is

considered to run from November to April (Ash et al. 2002), although its duration is highly

variable (see Figure 3.2. for mean monthly rainfall).

The experimental design was a randomized complete block with four replicates of each

treatment at each site. At Dalrymple, plots located on the levee bank were considered one

block. The other plots were blocked on the basis of percentage C. ciliaris cover, estimated as

described below. At Moorrinya, plots were blocked on the basis of percentage C. ciliaris

cover although one high C. ciliaris cover plot (plot 10) was incorrectly allocated to the low

C. ciliaris cover block. For most analyses, blocking was not significant and block effects

were ignored unless stated otherwise.

3.3.3. PRE-FIRE MEASUREMENTS

Prior to the imposition of burning treatments the plant species composition of plots was

recorded. Immediately pre-fire, fuel biomass and moisture content, air temperature and

relative humidity were determined for each fire.

3.3.3.1. Plant species composition

Plant species composition and abundance were assessed in April 1999 at Dalrymple and in

April and June 1999 at Moorrinya. Each plot was searched for 10-15 minutes and the species

found recorded. The height (estimated) and circumference at breast height (measured) of

trees and shrubs were also recorded. Twenty quadrats (1 m2) per plot were assessed in detail.

Quadrats were located by throwing a 1 m by 1 m wire frame from five positions along each

side of the plot. In each quadrat, the species present and their visually estimated percentage

91

cover (the percentage of the total vegetative cover in the quadrat) were recorded. Plant

specimens were collected for identification of unknown species and to produce a herbarium

collection.

3.3.3.2. Fuel biomass and moisture content

Fuel biomass and moisture content were estimated immediately prior to burning at each site.

Herbaceous biomass was estimated using a ‘standing crop disc’. This apparatus consists of a

metal rod marked with a scale and a metal disc with a central hole through which the rod is

inserted. The rod is held vertically with one end resting on the ground. The disc is held at the

top of the rod at a set position with its flat surface parallel to the ground. It is then allowed to

drop and the position on the rod at which the disc comes to rest is recorded. The greater the

amount of biomass under the disc, the higher it sits in relation to the rod and the higher the

‘score’. Although this method works best in continuous swards, random sampling ensured a

representative measure of biomass (bare ground was also sampled). The method allowed

rapid, non-destructive sampling providing a greater sample size than would have been

possible with destructive sampling. Eighty standing crop disc scores were randomly

collected at Dalrymple (60 within the main experimental area and 20 within the levee bank

area). Sixty standing crop disc scores were collected within the experimental area at

Moorrinya. An additional 10 disc score measurements were made at each site to calibrate

scores with biomass. After each of these 10 scores was recorded, all herbaceous material

under the disc (0.5 m2 quadrat) was cut at ground level. These samples were weighed after

being dried at 70oC. The relationship between score and dry weight of herbaceous material

(kg/ha) was determined by regression analysis.

Fuel moisture content was determined by collecting biomass samples around, but not within,

the experimental plots. At Dalrymple, 10 and five samples were collected around the main

and levee experimental areas respectively. Ten samples were collected at Moorrinya. Each

sample of 200-300 g of plant material was cut at ground level and placed in an air-tight

plastic bag. Samples were weighed fresh and reweighed after drying at 70oC.

3.3.3.3. Weather conditions

Temperature and relative humidity were measured with a hygrometer immediately prior to

lighting the fires.

92

3.3.4. FIRE CONDITIONS

Fire management resources dictated that plots allocated to burning treatments were burnt one

at a time. One measurement of fire speed was made in each plot using two markers. Markers

were placed in the direction of the fire, one about 10 m from the other and the distance and

time taken to burn between the markers were recorded. Fire intensity (I) for each plot,

defined as the heat released per metre of fire front (kW/m) was calculated using the

relationship:

I = Hωr

where H is the heat yield of the fuel burnt (the heat yield of herbaceous fuels is

approximately 17 kJ/kg (Byram 1959), ω is the amount of fuel consumed (kg/m2) and r is the

rate of spread (m/sec) (Cheney and Sullivan 1997).

The fires were generally slow moving and of low intensity, creating patchy burns with some

biomass remaining either unburnt or scorched in most plots. However, the late dry season

fires at Moorrinya were more intense, reducing most biomass to ash. A summary of the

conditions on the day of the fire and fire characteristics is given in Table 3.1. Fire intensities

were probably over-estimated since not all the fuel was burnt.

3.3.4.1. Early dry season fires

At Dalrymple on the 15th June 1999 conditions were cool (23oC at 1 pm) and relative

humidity was low (30% at 1 pm). Breezes were light and changed direction often. Plots were

burnt by lighting back burns before lighting head fires. The first fire was lit around 1.20 pm

and all fires were completed by about 3.30 pm. The fires produced patchy burns leaving

some plants scorched or unburnt. Estimated fire speeds ranged from 0.05 m/sec to 0.29

m/sec and estimated fire intensity ranged from 441 kW/m (plot 10) to 3030 kW/m (plot 1)

(Table 3.1).

At Moorrinya on the 10th June conditions were cool (23.5oC at 12.30 pm). Relative humidity

was 57% at 10 am and had declined to 48% by 12.30 pm. The prevailing breeze was from

the south-east. Plots were burnt with head fires lit from the south-east corners. The first fire

was lit around 11.15 am and all fires were completed by about 2.00 pm. The fires produced

patchy burns because of variability in fuel levels: the plots contained areas with little or no

biomass present. Some spot burning was carried out but kept to a minimum to simulate a

natural burn. Estimated fire speeds ranged from 0.07 to 0.12 m/sec and estimated fire

93

intensities ranged from 525 to 1125 kW/m (Table 3.1).

3.3.4.2. Late dry season fires

On the 12th November 1999, conditions at Dalrymple were warm and humid. The

temperature was 28oC and relative humidity 42% at 1.20 pm. Breezes were light and

changed direction often. The first fire was lit around 1.20 pm and all fires were completed by

about 3.00 pm. The fires generally produced very patchy burns leaving many plants scorched

or unburnt. Estimated fire speeds ranged from 0.07 m/sec to 0.22 m/sec and estimated fire

intensities ranged from 326 to 1619 kW/m (Table 3.1).

On the 17th of November 1999, conditions at Moorrinya were dry and gusty. Relative

humidity was 35% and the temperature 34oC at 11.00 am. The first fire was lit around 11.00

am and all fires were completed by noon. All fires burnt strongly leaving relatively few

patches unburnt and reducing most above ground biomass to ash. Estimated fire speeds

ranged from 0.07 m/sec to 0.31 m/sec and estimated fire intensities ranged from 462 to 2047

kW/m (Table 3.1).

Table 3.1. Summary of conditions for early and late dry season fires at Dalrymple and

Moorrinya National Parks. Values are means ± SE.

Site Date Fuel loada

(kg/ha) Fuel moisture

(%) Fire speed

(m/sec) Fire intensity

(kW/m) Dalrymple 15/6/99 6150 ± 238b

5180 ± 241c33 ± 1.3 0.14 ± 0.075 1400 ± 819

12/11/99 4330 ± 320b

2950 ± 720c26 ± 1.6 0.12±0.036 820 ± 285

Moorrinya 10/6/99 4410 ± 245 25 ± 2.1 0.12 ± 0.017 860 ± 125 17/11/99 3880 ± 272 28 ± 2.1 0.22 ± 0.052 1490 ± 351 a. Mean biomass was estimated from standing crop scores using the following relationships

(regression analyses performed using Genstat (2001):

Dalrymple: Early dry season fire biomass = 371+36*1.11score (Adj R2 = 0.85)

Dalrymple: Late dry season fire biomass = 59.38*Score0.91 (Adj R2 = 0.88)

Moorrinya: Early dry season fire biomass = 60.04*Score0.76 (Adj R2 = 0.97)

Moorrinya: Late dry season fire biomass = 23.55*Score1.09 (Adj R2 = 0.97)

b. Mean fuel load in main experimental area.

c. Mean fuel load in levee bank area.

94

3.3.5. LOCAL RAINFALL

Detailed climate data were not available for the experimental sites. Rainfall data for Fletcher

View station, 11 km south-east of Dalrymple National Park are given in Figure 3.2a. Rainfall

data were available for Moorrinya National Park from January to August 1999 and from

October 2000 to June 2001. These data and data from Uanda Station, about 14 km south-

west of Moorrinya, for the missing months are given in Figure 3.2b.

0

50

100

150

200

250

300

350

400

Mon

thly

pre

cipi

tatio

n (m

m)

0

50

100

150

200

250

300

350

400

Mon

thly

pre

cipi

tatio

n (m

m)

Mon

thly

pre

cipi

tatio

n (m

m)

Figure 3

Fletcher

(except

Station i

stations

Parks an

(a)

J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J

(b)

J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J1998 1999 2000 2001

.2. Monthly precipitation (bars) recorded from June 1998 to June 2001 at (a)

View Station near Dalrymple National Park and (b) Moorrinya National Park

from September 1999 to September 2000 when precipitation for nearby Uanda

s given). Line graphs show monthly rainfall averages for Fletcher View and Uanda

respectively (Commonwealth Bureau of Meteorology, unpublished; Queensland

d Wildlife Service, unpublished).

95

3.4. CAN FIRE KILL CENCHRUS CILIARIS?

3.4.1. INTRODUCTION

The aim of weed management is to limit or reduce the abundance of invasive species with

minimum negative impacts of the weed or the treatment on the establishment and vigour of

desirable species. A key factor determining the efficacy of fire as a tool for manipulating the

composition of C. ciliaris-dominated grassland is the effect of fire on C. ciliaris plant

persistence.

Competition from established plants is a major factor preventing plant establishment in

perennial grasslands (Cook 1980). Consequently, the longevity of existing plants is a

significant factor influencing community dynamics. Astrebla plants may live for 23 years

(Orr and Holmes 1984) and C. ciliaris plants are also considered to be long-lived (Brown

1985). Habitat disturbances that reduce or remove established plant competition are

important since they create gaps that could provide ‘safe sites’ (Harper 1977) for

germination and seedling establishment (Cheplick 1998) (section 3.2.3.5). Fire is an

important agent in structuring communities since the openings it creates provide the potential

for vegetation change (Bond and van Wilgen 1996). By removing perennial vegetation, fire

may free up resources and create conditions more conducive to germination (Cheplick and

Quinn 1988) and establishment and many studies report increased seedling emergence and/or

establishment in grasslands after fire (Shaw 1957; Christensen and Muller 1975; Williams et

al. 2003b) (section 3.2.2.5). In C. ciliaris-dominated grasslands, the effects of fire on other

ecosystem components and processes, such as seed banks and flowering, are likely be

irrelevant in promoting compositional change unless C. ciliaris plants can be removed and

resources made available to support the establishment of new individuals.

Can fire be used to kill C. ciliaris plants? In this study the effects of season of burning on the

survival of C. ciliaris were investigated. At Moorrinya the effects of season of burning on

the survival of the dominant native Astrebla spp. were also investigated. At Dalrymple,

native species such as Heteropogon contortus were not abundant enough to include in the

study. Season of burning is important in determining the extent of fire-induced plant

mortality (Trollope 1997) due to seasonal differences in fire and plant characteristics (section

3.2.3.1). Late dry season fires are generally more intense than early dry season fires

(Williams et al. 1997) and, since the level of plant mortality is positively related to fire

intensity (Williams J.R. 1995), late dry season fires may result in higher plant mortality

96

(Lonsdale and Braithwaite 1991; Williams et al. 1997). The timing of fire in relation to plant

phenology also influences the degree of damage caused by fire (West 1965; Tainton 1981).

Tissue moisture content and the location of carbohydrate stores (roots or shoots) varies

seasonally and affects both plant susceptibility to fire and recovery after fire (Bond and van

Wilgen 1996). Actively growing species may be highly susceptible to fire while dormant

species are relatively unaffected (Daubenmire 1968; Bond and van Wilgen 1996).

Consequently, although late dry season fires may be more intense than early dry season fires,

early dry season fires may cause more injury if plants are still actively growing.

3.4.2. METHODS

3.4.2.1. Field set-up

Cenchrus ciliaris plants in the experimental plots at Dalrymple and C. ciliaris and Astrebla

spp. plants in the experimental plots at Moorrinya were tagged in April 1999 using numbered

metal pegs pushed into the soil beside each plant. At Dalrymple, 10 large and 10 small C.

ciliaris plants were tagged per plot. Plants were considered small if their basal area was less

than 50 cm2. Plots were searched to provide 10 pairs of plants where possible (each pair

being a large and small tussock within 1 m of each other). The growth habit of C. ciliaris

was such that it was often difficult to determine the origin of tillers and thus identify

individual tussocks. Plants difficult to distinguish as individual tussocks were avoided. At

Moorrinya, 20 C. ciliaris and 20 Astrebla plants were tagged per plot. Most Astrebla plants

were A. squarrosa although some individuals of A. pectinata were included. Plants were

chosen to include a range of sizes. The difficulty in determining individual tussocks was

even greater at this site. Clearly identifiable C. ciliaris and Astrebla tussocks were selected

where possible but in some cases the extent of individual tussocks was uncertain.

The basal area of each tagged plant was estimated as:

Basal area (cm) = π * diameter 1(cm) * diameter 2 (cm)/4

where the diameters were the longest and shortest lengths of the tussock base. Tussock base

measurements were made prior to the imposition of burning treatments.

The post-fire status (dead or alive) of tagged plants was recorded in January and June 2000

at Dalrymple and in February and August 2000 at Moorrinya. The results were similar

97

between sampling times and only the latter results are reported.

3.4.2.2. Statistical analyses

Burning treatment effects on the percentage of dead small plants at Dalrymple were

investigated using one-way ANOVA. Data on large plants were not analysed since only one

large plant died. The percentages of dead plants at Moorrinya were analysed using a split-

plot design with burning treatments as main plots and species as subplots after checking for

homogeneity of error variances. Data were arcsin transformed prior to analysis to normalize

the distribution of residuals where necessary. In these cases, the reported means have been

back-transformed. Means were compared using the protected LSD test at the 5% significance

level. Basal areas were compared between small and large plants at Dalrymple and between

C. ciliaris and Astrebla plants at Moorrinya using t-tests assuming unequal variances.

Regression analysis was used to investigate relationships between plot fire intensity and the

percentage of dead plants per plot. All analyses were conducted using Genstat (2001).

3.4.3. RESULTS

Plant mortality was low at both sites with less than 6% of tagged plants dying over the

monitoring period. Despite this, burning treatment effects were detected.

At Dalrymple, the mortality of small plants was greatest in the early dry season burn

treatment (P < 0.05) where 24% of plants died compared with less than 3% in other

treatments (Table 3.2). In contrast only one large tagged plant died at the site. At Moorrinya,

C. ciliaris and Astrebla mortalities were greatest in the late dry season burn treatment (Table

3.2). No differences in mortality between C. ciliaris and Astrebla plants were detected

(P>0.05).

Generally, the plants that died had smaller basal areas than average. Of the dead small plants

at Dalrymple, all but one had smaller basal areas than the average for that size class. The

basal area of small plants ranged from 0.1-44.0 cm2 (mean ± SE = 12.6 ± 1.03) while the

basal area of killed small plants ranged from 0.1-15.7 cm2 (mean ± SE = 3.4 ± 1.39 cm2).

The large plant that died was also smaller than average (77.8 cm2) (Table 3.2). At

Moorrinya, the average size of dead plants (72 ± 37.9 cm2 and 72 ± 30.2 cm2 for C. ciliaris

98

and Astrebla plants respectively) was less than the average for these species (Table 3.2).

Only two of the 15 dead C. ciliaris plants and two of the 10 dead Astrebla plants were larger

than average.

At both sites over 50% of tagged plants had basal areas of 100 cm2 or less. The size class

frequency distribution of tagged plants is given in Figure 3.3. At Moorrinya, a few very large

C. ciliaris plants were tagged and mean basal area of C. ciliaris plants was significantly

greater (P < 0.05) than that of Astrebla plants.

Table 3.2. Mean basal area (cm2) and mortality (%) of tagged plants at Dalrymple and

Moorrinya. Different lower case letters denote significantly different (P < 0.05) percentage

mortality within plant size classes (Dalrymple) and within species (Moorrinya).

Site Plant type Plant mortality

(% of tagged plants that died)

basal area

(mean ± SE)

(cm2) Control Early burn Late burn

Dalrymple

Small 13 ± 1.0 *3a *24b *1a

Large 180 ± 13.1 0 3 0

Moorrinya

C. ciliaris 244 ± 22.6 3a 3a 14b

Astrebla 137 ± 10.0 0a 0a 14b

* Back-transformed means

99

0

10

20

30

40

50

60

70 (a)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

0

10

20

30

40

50

60

70

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

(b)

Per

cent

age

of ta

gged

pla

nts

0

10

20

30

40

50

60

70

Figure 3.3.

(a) C. ciliar

at Moorriny

Linear relat

were signifi

distribution

1 2(c)

Basal area (cm2) 221 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 210 500 1000 1500 2000

Percentage of tagged plants in basal area size classes (increments of 100 cm2).

is plants at Dalrymple, (b) C. ciliaris plants at Moorrinya and (c) Astrebla plants

a.

ionships between plot fire intensity and the percentage of dead tagged plants

cant (P < 0.05) for both sites. However, individual plots had high leverage and

of the data prevented valid assessment of these relationships (Figure 3.4).

100

0

10

20

30

40

50

60

0 500 1000 1500

Fire Intensity

Per

cent

age

dead

pla

nts

(%)

a

1e

8

Figure 3.4. Relationships between plot fire intensi

died/plot at (♦) Dalrymple and (□) Moorrinya. Regre

3.4.4. DISCUSSION

Fire-induced plant mortality is related to fire intensi

mortality with higher intensity fires (Williams J.R.

2001). In this study also, more plants were killed in

relationship between fire intensity and season of bu

Moorrinya, the mean intensity of late dry season

season fires while at Dalrymple, the reverse was fou

late dry season fires at Dalrymple were due to lower

the time of burning.

Although greater plant mortality was associated with

season of burning, it was not possible to determine a

and mortality using the individual plot data. A more

is needed to validly assess fire intensity-plant morta

recognize that plot-level fire intensities may not

conditions that are critical in determining the fate o

load and only one fire speed estimate per plot we

Dead%= 0.007*Fire intensity-0.08Adj. R2=0.53

Dead% = 0.01*Fire intensity +1.5
Adj. R2=0.5
2000

(kW/m

ty and

ssion li

ty and

1995;

the hig

rning d

fires w

nd. Th

fuel lo

higher

clear r

even d

lity rel

adequa

f indiv

re used

Dalrympl

Moorriny

2500 3000 3500

)

percentage of tagged plants that

nes are shown.

many studies report greater plant

Lippicott 2000; D’Antonio et al.

her intensity fires. However, the

iffered between the two sites. At

as higher than that of early dry

e unexpectedly low intensities of

ads and the weather conditions at

mean fire intensity, regardless of

elationship between fire intensity

istribution of plot fire intensities

ationships. It is also important to

tely describe the fine scale fire

idual plants. An average site fuel

to calculate plot fire intensity.

101

However, the fire intensities experienced by individual tussocks clearly varied widely. Due

to the patchy nature of the fires some plants were reduced to ash while others remained

unburnt. In addition, fire residence time may be more biologically important than fire

intensity (Walker et al. 1981; Noble and Grice 2002) (section 3.2.3.1).

Fire intensity probably explains the reversal of season of burning effects between the two

sites. However, differences in the growing season and phenological condition of plants

between sites may offer an alternative explanation for the increased early dry season

mortality at Dalrymple. Since plants with higher moisture contents are killed at lower

temperatures (Bond and van Wilgen 1996), actively growing plants are more vulnerable to

fire (Daubenmire 1968; Bond and van Wilgen 1996). If C. ciliaris plants at Dalrymple were

still growing at the time of the early dry season fires, they may have been more vulnerable to

fire than dormant plants. It is likely that plants were dormant by the time of the late dry

season fires. A shorter growing season at Moorrinya may have resulted in plants being

dormant at this site at the time of the early dry season fires. Although it is not possible to

estimate the duration of the growing seasons, at the time of the early dry season fires at

Dalrymple, fuel moisture contents were higher (33 ± 1.3%) than those at the late dry season

fires (26 ± 1.6%) at this site and higher than those at any fires at Moorrinya (25 ± 2.1 and 28

± 2.1% moisture for early and late dry season fires respectively) (section 3.3.4). This may

indicate that the vegetation at Dalrymple had not senesced to the degree it had at Moorrinya.

It is unlikely that the amount of biomass removed in these fires had much effect on

establishment site availability. Although burning resulted in C. ciliaris and Astrebla

mortality, the numbers of plants killed were low and those killed tended to be smaller than

average. The greater susceptibility of small plants was not surprising since smaller plants are

brought to lethal temperatures more quickly than larger plants (Bryam 1948) and other

studies have found fire-induced mortality is greater for smaller individuals (refer Bond and

van Wilgen 1996; Grice 1997) (section 3.2.3.2). The death of the few small plants in this

study would have been insignificant in opening up establishment sites in these grasslands.

Although the removal of large individuals is more effective in creating establishment sites,

the susceptibility of small plants to fire is important since it is likely that a large proportion

of individuals in these populations are small. Although the size distribution of tagged plants

may not accurately represent the plant size distributions of these populations (since tagged

plants were not selected randomly), the actual size class distributions are probably skewed,

as found for the tagged plants. If so, and if plant size is an important determinant of

susceptibility to fire, it appears that many individuals in these populations are susceptible to

102

relatively low intensity fires. Ninety percent of tagged Astrebla plants and 60% of tagged C.

ciliaris plants were of the same size or smaller than the largest killed plants. The small

numbers of plants actually killed probably reflects the patchy nature of the burns in which

some plants were reduced to ash while others were barely singed. A more uniform burn is

likely to cause more deaths and create more significant gaps. Higher intensity fires are likely

to be even more effective if larger individuals are also killed.

Co-occurring species may respond very differently to the same fire (Daubenmire 1968). For

example, in burning experiments in Hawaiian woodland, Schizachyrium condensatum

tolerated fire with 94% and 40% plant survival after low and high intensity fires while

Melinis minutiflora was greatly affected with 30% and 0% plant survival after the same fires

(D’Antonio et al. 2001). Such responses may occur where species differ in condition (size

and vigour), morphology, and/or life strategy (section 3.2.3.2). For example, differences in

plant responses to fire may be related to plant architecture and the degree to which

meristematic sites are protected from heat injury (Tainton 1981; Walker et al. 1983). The

more loosely packed tillers of Heteropogon contortus may make it more susceptible to heat

damage from fire than the dense internally damp tussocks of Themeda triandra which

effectively insulate the apices against heat damage (Walker et al. 1981). However, the

relative responses of species to fire may change over time and it appears that T. triandra may

be more susceptible to fire later in the dry season since its tussocks produce significantly

higher temperatures than H. contortus tussocks (refer Walker et al. 1983).

Different responses to fire by C. ciliaris and local native species are of particular interest.

However, in this study it was not possible to monitor fire effects on native species other than

Astrebla at Moorrinya. Further work is required to investigate fire-induced mortality of C.

ciliaris compared with that of native species. Are C. ciliaris plants susceptible to the

strategic use of fire? The removal of above ground biomass by fire is likely to have less

impact on C. ciliaris than on some native species. Defoliation experiments have found that,

while defoliation reduces root biomass, C. ciliaris maintains a relatively high root:shoot ratio

compared with some native species (Brown 1985) and its persistence under heavy grazing is

testimony to its ability to recover after defoliation. However, high intensity fires may kill C.

ciliaris plants. Fatal fire intensities may be easier to achieve in C. ciliaris patches since C.

ciliaris is associated with greater biomass production than native species (Humphries et al.

1991; Latz 1991), although this is not always the case (section 2.2.3). Differences in plant

phenology may be exploited. Cenchrus ciliaris has been reported to remain green and cure

later than native species (Cavaya 1988; Humphries et al. 1991). This may make it more

susceptible to early dry season burns than native species which senesce earlier.

103

3.5. DOES FIRE INCREASE PLANT NUTRIENT AVAILABILITY IN CENCHRUS

CILIARIS-DOMINATED GRASSLAND?

3.5.1. INTRODUCTION

Fire alters plant community productivity, composition and structure via direct effects on

plant growth, survival and reproduction. However, fire-induced changes in the

environmental conditions experienced by plants may also result in significant vegetation

change. Fire alters the availability of light, water and nutrients and causes changes in

microbial and animal populations (section 3.2.2.4). Although plants respond to the combined

influence of these and other post-fire conditions, changes in one factor can be a significant

determinant of vegetation change. Fire can rapidly alter the amount, distribution and form of

plant nutrients (Raison 1979) and fire-altered nutrient status may have major impacts on

plant communities. Enhanced plant growth (Christensen and Muller 1975; Scanlan and

O’Rourke 1982; Hulbert 1988) and changes in species composition (Daubenmire 1968;

D’Antonio et al. 2001) have been attributed to post-fire increases in nutrient resources.

Fire directly affects plant nutrient availability by incinerating organic matter and by heating

the soil (Raison 1979). It indirectly affects nutrient availability via effects on the activity of

soil biota (Christensen and Muller 1975), water movement and erosion processes (Kellman

et al. 1985). Changes to soil properties may also result from fire effects on plant community

composition, structure and productivity (Raison 1979).

Nutrients are released from burnt organic matter and nutrient availability is commonly,

though not universally, increased by fire due to ash added to the soil surface (Bond and van

Wilgen 1996). Ash contains large quantities of various nutrients in mineral form as well as a

readily available soluble, organic reservoir of these materials (Christensen and Muller 1975).

Fire also alters soil chemistry via heating effects (Raison 1979). For example, soil

phosphorus and nitrogen fractions may be altered by heat (refer Humphries and Craig 1981).

As well as changes in the amount and form of nutrients, nutrient concentrations may be

considerably more variable after fire due to variations in fire intensity and the uneven

distribution of ash (Christensen 1985). The effects of ash and heat on soil chemistry and

plant responses to these effects are reviewed by Raison (1979).

Not all the nutrients from burnt biomass are retained in the soil. Some nutrients, particularly

nitrogen and sulphur, may by lost through volatilisation and nutrients may be removed as

particulates via air and water movement (Kellman et al. 1985). Although it is often

104

considered that most of the N and S in biomass is volatilized during fires and most of the

phosphorus, potassium, calcium and magnesium is deposited in ash, P, K and Mg can be

volatilized during fires and all elements can be transferred to the atmosphere as particulates

(Cook 1992). The magnitude of nutrient losses due to fire depends on fire intensity, flame

dimensions and temperature profiles, the spatial pattern of these properties across the

landscape as well as fire frequency (Cook 1992). The proportion of nutrients in volatile

versus particulate states depends on the vapourization temperature of the elements involved

and the fire intensity (Raison et al. 1985). The more intense the fire, the greater the

volatilization losses (Bond and van Wilgen 1996; Mack et al. 2001). Volatilization losses

from soil organic matter are likely to be low in savannas because soil temperatures and

intensities of savanna fires are relatively low and the build up of organic matter between fires

is also relative low (Cook 1992). However, volatilization losses from aboveground biomass

can be significant. Nitrogen losses due to volatilisation have been found in north American

tallgrass prairies (Knapp and Seastedt 1986; Turner et al. 1997) and savannas in the Northern

Territory, Australia (Cook 1992; 1994).

Nutrients may also be lost via air and water movement. Nutrients transported into the air as

particulates may be re-deposited. Since the movement and redistribution of ash by wind is

patchy (Cook 1994) and material can settle out at the site or many kilometres away, the

assessment of whether such material is lost or re-distributed depends on the scale considered.

Fire converts nutrients to readily soluble forms, increasing the potential for losses via

leaching. However, leaching losses may be insignificant where there is rapid uptake by

plants (Cook 1992) or where nutrients are rapidly immobilized in the soil after fire (Kellman

et al. 1985). Removal of nutrients in runoff can be a major cause of nutrient loss (Kellman et

al. 1985). The nature and season of the burn, the interval between burning, topography, soil

type and the timing and intensity of rainfall influence fire effects on run-off and erosion

(West 1965). Post-fire nutrient losses via runoff have been reported in Australian savannas

(Douglas et al. 1996). However, while runoff and soil loss is related to soil cover, soil

structure also has a significant influence and soil losses after fire may be less than expected

if infiltration rates are high (McIvor et al. 1995).

In the longer term, nutrient availability after fire may be affected by increased soil

temperatures and altered soil moisture and pH. The removal of surface cover by fire results

in increased air and soil temperatures and greater temperature variability (refer Christensen

1985; Whelan 1995) whereas moisture conditions may increase or decrease after fire (section

3.2.2.4). Soil pH generally increases after fire due to the addition of ash (Daubenmire 1968;

Christensen and Muller 1975). These changes may directly affect nutrient availability. They

105

may also effect nutrient availability indirectly via effects on microbial activity, and

consequently, nutrient transformations. Rates of mineralization may be increased due to

changes in pH, reductions in C:N ratios (refer Van de Vijver et al. 1999) and increased

temperatures (Daubenmire 1968). Nutrient cycling may also be affected by fire-induced

changes in plant species composition and both increases (refer Daubenmire 1968) and

decreases (Mack et al. 2001) in nitrogen fixation have been attributed to species changes.

Nutrient availability is influenced by the degree of uptake by plants and the nutrients

available to individual plants may be affected by the altered competitive relationships that

exist after fire.

Soil fertility is fundamental in controlling the functioning of plant communities and an

understanding of the effects of various fire regimes on soils is needed to properly evaluate

fire-management options (Raison 1980). Plant competitive interactions are influenced by

nutrient conditions (Howden 1988) and exotic species such as C. ciliaris may be favoured by

post-fire flushes of nutrients. The aim of this study was to investigate the effects of season of

burning on plant nutrient availability in C. ciliaris-dominated grassland. I expected that, in

general, fire would enhance nutrient availability. However, season of burning effects were

more difficult to predict. Season of burning effects on soil nutrients depend on the timing of

fire in relation to vegetation condition and on fire characteristics, such as extent and

intensity. Late dry season, high intensity fires may convert more biomass to ash than milder,

early dry season fires, increasing the supply of readily available nutrients. However,

volatilization losses will be greater in more intense fires. In addition, nutrient concentrations

may be lower in late dry season fuel, and consequently late dry season ash, since plants have

relocated nutrients to roots by this time (Norman 1963).

3.5.2. METHODS

3.5.2.1. Soil collection and processing

A bioassay technique was used to determine the surface soil nutrient availability of all

experimental plots at the start of the growing season. Soil samples were collected from

unburnt (control) plots and from plots burnt in the early or late dry at Dalrymple and

Moorrinya after the late dry season fires (section 3.3.1). Samples were collected on the day

of the fires at Moorrinya. Due to time constraints only block one was sampled immediately

106

after the fires at Dalrymple and the other blocks were sampled 20 days later. Six surface soil

samples, approximately 20 cm by 20 cm by 5 cm deep, were dug from random positions

within the central 18 m by 18 m zone of each plot. Samples were collected beside plant bases

when the random sampling position fell on a plant base. Where present, surface ash was

included in the samples. The six samples were bulked, providing one sample per plot.

Samples were air-dried before being processed with a concrete roller to crush soil

aggregates. Samples were then sieved (to pass a 6.7 mm mesh) and mixed thoroughly. While

in storage some samples were accidentally wet. To insure samples received similar

processing, all soils were wet, re-dried and re-sieved.

3.5.2.2. Pot set up

The experiment was set up in a glasshouse at Davies Laboratory, Townsville (19o15’S,

146o45’E) in September 2000. Three pots (approximately 13 cm diameter, 14 cm high) of

each sample were prepared using either 1600 g Dalrymple soil/pot or 1500 g Moorrinya

soil/pot. Since the soils from the two sites differed in bulk density, different weights of soil

were used to achieve a similar volume. Pots were free draining but were lined with gauze to

prevent any soil loss with watering. Pots were arranged in four blocks, the allocation of pots

to blocks corresponding to the blocking of the field plots. The pots were re-randomized

within blocks at frequent intervals and kept well watered.

Twelve forage sorghum (Sorghum bicolor (L.) Moench, variety ‘Sudenensis’) seeds were

sown in each pot. This species was chosen because it is highly responsive to soil nutrients

and re-shoots readily after decapitation (A. Noble pers. comm.). During the first week

additional pre-germinated seed was added to achieve uniform emergence. Seven days after

sowing emerging seedlings were thinned to three plants per pot and a layer of black plastic

beads (1 cm deep) was added to the soil surface to prevent algal growth.

The experiment was run for 31 weeks, from September 2000 to April 2001. Over this period

above ground biomass was harvested five times at approximately six week intervals. For

each pot, all plants were cut at the soil surface, placed in a paper bag and dried at 65oC

before being weighed. After the fourth harvest, plants failed to re-shoot in some pots and no

plants re-shot after the fifth harvest. The cumulative biomass produced from each pot was

calculated for the five harvest periods.

107

3.5.2.3. Statistical analysis

Total cumulative biomass data for each harvest period were analysed separately using one-

way ANOVA procedures in Genstat (2001). Data from each site were analysed separately

and since block effects were not significant ANOVA without blocking results are reported.

Results for each harvest were similar and only the final (total) biomass data are reported.

Relationships between plot fire intensity and sorghum biomass were investigated using linear

regression, also in Genstat (2001). While the biomass produced from pots containing

Dalrymple soil could not be directly compared with that from pots containing Moorrinya soil

because different weights of soil were used, the amount of biomass produced per gram of

these soils was calculated.

3.5.3. RESULTS

No burning treatment effects on above-ground sorghum biomass were detected for either site

(P > 0.05) (Figure 3.5). Plants grown in Dalrymple soil produced more biomass per gram of

soil (3.2 ± 0.13 mg biomass/g soil) than plants grown in Moorrinya soil (2.2 ± 0.06 mg

biomass/g soil).

Figure 3.5. Mean (± SE) total above-ground biomass (g/pot) of sorghum plants grown in soil

from ( ) control (no burn), ( ) early dry season burn and ( ) late dry season burn plots at

Dalrymple and Moorrinya.

0

1

2

3

4

5

6

7

Dalrymple Moorrinya

Mea

n to

tal b

iom

ass

(g/p

ot)

108

There was a significant linear relationship between plot fire intensity and sorghum biomass

for Dalrymple (P < 0.05; adjusted R2 = 0.57) (Figure 3.6). However, no relationship was

detected when the high intensity plot, which had high leverage, was omitted from the

analysis. No relationship between plot fire intensity and sorghum biomass was found for

Moorrinya (P > 0.05) (Figure 3.6).

0

1

2

3

4

5

6

7

8

9

0 500 1000 1500 2000 2500 3000 3500

Fire intensity (kW/m)

Bio

mas

s (g

/pot

)

Figure 3.6. Relationship between plot fire intensity (kW/m) and mean total above ground

sorghum biomass (g/pot) for (♦) Dalrymple and ( ) Moorrinya. Regression line for

Dalrymple shown (Biomass = Plot fire intensity*0.001 + 4.4).

3.5.4. DISCUSSION

Development of fire management regimes requires a consideration of nutrient dynamics if

conservation and productivity values are to be maintained (Cook 1992). The single fires

reported in this study appeared to have little effect on soil nutrient availability. Other studies

have also failed to detect significant differences in nutrient availability in

grasslands/savannas due to fire (Shaw 1957; Harrington and Ross 1974).

109

Fire effects on nutrient availability may vary between early and late dry season fires since

the season of burning is associated with differences in fire intensity (Lonsdale and

Braithwaite 1991; Williams et al. 1997) and differences in fire intensity may result in

different effects on nutrient status (refer Christensen and Abbott 1989). However, in this

study fire intensity was not related to season of burning in the usual way since the mean

intensity of late dry season fires was lower than that for early dry season fires at Dalrymple.

In addition, nutrient availability appeared to be unaffected by fire intensity: although the

higher intensity fires converted more biomass to ash, there was no convincing evidence that

nutrient availability was related to fire intensity.

Nutrient conversion and loss processes due to burning in these grasslands remain unclear.

Although burning may enhance nutrient availability, some nutrients may be lost via

volatilisation or removal from the site by leaching, run-off and the erosive action of wind

(Raison 1980). Volatilisation losses of N have been reported from savanna fires (Cook

1992;1994). Also, while about three-quarters of the ash stays on the ground, the remainder is

carried up in the smoke plume to be deposited many kilometres away (Cook 2001) and

losses of ash by water erosion may be important in certain landscapes after late dry season

fires (Cook 1992). The generally patchy, low intensity fires in this study did not convert all

the standing biomass to ash and volatilization losses were probably minimal. Grassland fires

are generally of lower intensity than forest fires (Williams J. R. 1995) and the quantity of

nutrients released by burning savannas is only a fraction of that released in forest situations

(Nye 1959). Daubenmire (1968) concluded that the ash produced by grass fires was so

scanty compared to that produced by forest fires that it seemed negligible. It is likely that the

fires in this study produced insufficient ash to cause detectable changes in nutrient

availability. Alternatively, given the patchy distribution of ash, it is possible that insufficient

soil samples were taken to adequately sample the burnt plots.

It is unclear how much of the ash produced remained on site. Air and water-borne

particulates may be re-deposited within a site, representing spatial re-distribution rather than

loss. Generally the smoke plumes of these fires were small and did not travel far. Although

there was considerable time (about 20 weeks) for ash to be blown from the early dry season

burnt plots, there were plenty of obstructions to trap and retain wind blown material. Rainfall

prior to soil sampling may have promoted nutrient uptake by plants or microbes or resulted

in loss in runoff water. There was significant rainfall at Moorrinya (85mm) two weeks prior

to the late dry season fires. However, rainfall details for the experimental sites are not

available and it is not possible to estimate pre-sampling nutrient losses.

110

Bioassay techniques have been used by other investigators to determine the effects of fire on

plant nutrients (Christensen and Muller 1975). This method has cost advantages over

chemical determination of soil nutrients plus it provides a clear demonstration of how plants

will respond to the total nutrient profile. However, the method used here does not permit

investigation of the fate of specific nutrients. Fire may have differential effects on nutrients.

However, since plant growth is determined by the most limiting factor, increases in one

nutrient may not be detected if growth is limited by another. In addition, while ash is a

supply of readily available nutrients, high nutrient concentrations in ash may be deleterious

to plant growth (Facelli and Kerrigan 1996). Therefore it is possible that the fires did affect

nutrients but that these effects were not detected. Difficulties in studying fire-nutrient

interactions are reviewed by Raison (1980).

The results of this study suggest that burning has little effect on soil nutrient availability in

these C. ciliaris-dominated grasslands. Compared with vegetation communities that

experience severe fires, the effects of grassland fires on nutrient availability may be

relatively small (refer Raison 1980). Also, the amounts of nutrients lost in any one fire are

generally small compared with the reserve in the soil and biomass (Raison 1980). In this

study, the effects of single, generally low intensity, patchy fires were investigated. However,

one-off fires are unlikely to be effective in achieving management goals and, in the context

of manipulating species composition, the effects of repeated burning on the nutrient

dynamics of these grasslands need to be determined. The cumulative effect of small but

persistent losses of nutrients from infertile soils may decrease site productivity and cause

degradation in the long term (Kellman et al. 1985; Cook 1994). Cenchrus ciliaris is favoured

by moderate to high fertility conditions (Humphries 1967; McIvor 1984). If burning

increases nutrient availability, C. ciliaris establishment may be favoured over that of native

species while a decline in nutrient availability may favour native species establishment. This

study has not shown any effect of single fires on nutrient availability.

111

3.6. LITTER: A HELP OR HINDRANCE TO SEEDLING EMERGENCE?

3.6.1. INTRODUCTION

Fire may promote changes in plant community composition by altering the availability of

sites suitable for the germination and establishment of new plants. The availability of such

‘safe sites’ (Harper 1977) depends on the germination and growth requirements of seeds and

seedlings and on the micro-site conditions they experience (section 3.2.3.5). The nature and

effects of the environment immediately surrounding seeds and seedlings are of critical

importance in determining the dynamics of plant populations and the composition of plant

communities (Fowler 1988). Litter influences the condition of seed/seedling micro-sites and

many studies report the effects of litter on germination and establishment patterns (Scanlan

and O’Rourke 1982; Enright and Lamont 1989; Facelli and Pickett 1991a; O’Connor 1991).

By altering the distribution and abundance of litter, fire may have significant impacts on

establishment patterns (Lane and Williams 1997). An understanding of the effects of litter on

germination and emergence may help predict the effects of fire on plant community

composition.

Litter affects plant communities by altering the chemical and physical conditions

experienced by plants. Litter alters the chemical environment by modifying the levels and

distribution of nutrients (Facelli and Pickett 1991a). Nutrient dynamics are influenced by the

composition of litter. For example, differences in nitrogen cycling associated with five

perennial grass species were attributed, in part, to differences in the nitrogen and lignin

content of their litter (Wedin and Tilman 1990). Nutrient dynamics are also influenced by

litter’s effects on environmental variables, such as temperature and pH, that regulate soil

biota and mineralization (Knapp and Seastedt 1986; Facelli and Pickett 1991a). Although

litter releases nutrients as it decomposes, it may result in reduced nutrient availability. For

example, Bromus tectorum litter was found to decrease inorganic nitrogen availability

because its low nitrogen content resulted in reduced nitrogen mineralization (Evans et al.

2001). In grasslands, accumulated litter may alter the chemical composition of rainfall

reaching the soil surface, decreasing nitrogen deposition (Knapp and Seastedt 1986). The

removal of litter and subsequent increases in soil temperature may promote organic matter

decomposition, increasing nutrient availability (refer Facelli and Pickett 1991a). The

leaching or decomposition of litter may also alter the chemical environment via phytotoxic

effects (see Rice 1979). However, the ecological role of such effects is debated (Barritt and

Facelli 2001) since there is little understanding of allelopathic processes in the field (Facelli

112

and Pickett 1991a).

Litter affects the physical environment by intercepting light and rain and by affecting the

transfer of heat and water, and probably gases, between the soil and atmosphere (Facelli and

Pickett 1991a). Litter reduces light levels (Facelli and Picket 1991b; Facelli and Kerrigan

1996) and can alter light quality (Facelli and Pickett 1991a). By intercepting solar radiation

and insulating the soil, litter also modifies soil temperatures (Evans and Young 1970) and

smaller diurnal ranges of surface and sub-surface temperatures have been found under litter

compared with bare soil (Evans and Young 1970). This moderation of temperatures may

affect plants directly, by promoting or inhibiting germination and/or protecting plants from

frost, and indirectly, via effects on nutrient dynamics (see Facelli and Pickett 1991a). While

soil temperatures may be reduced under litter (Facelli et al. 1999), leaf and air temperatures

may be increased due to reduced convective cooling (Knapp and Seastedt 1986). Litter

affects water dynamics by altering the exchange of water between the soil and atmosphere

(Facelli and Pickett 1991a). Moisture conditions under litter are generally more stable than

on bare soil since moisture depletion is less rapid (Evans and Young 1970). In grasslands,

litter is associated with increased water availability as it reduces rain drop impact (Bridge et

al. 1983), increases infiltration and reduces evaporation and run-off (see Facelli and Pickett

1991). However, in some circumstances litter reduces moisture availability by retaining

rainfall which is then lost by evaporation (Knapp and Seastedt 1986), and by increasing

runoff (Walsh and Voigt 1977). Litter may act as a physical barrier (Facelli and Pickett

1991a), impeding the movement of seeds and/or seedlings to or from the soil surface.

The effects of litter on plant development are complex. The environmental changes caused

by litter may affect plants directly, by influencing germination and establishment processes,

and indirectly, via effects on plant interactions (Facelli and Pickett 1991a; Facelli 1994),

herbivory (Facelli 1994) and disease (Facelli et al. 1999). The environmental changes

associated with litter depend on characteristics of the litter. Plant responses to litter-altered

conditions depend on characteristics of the plant species involved (Facelli and Kerrigan

1996). For example, characteristics such as the shape and size of seeds (Evans and Young

1970; Facelli and Pickett 1991a) and the shape of seedlings (Gross and Werner 1982)

influence their movement through litter. Plant physiological responses to light and

temperature also influence the effects of litter on germination and establishment (refer

Facelli and Kerrigan 1996). In addition, plant responses to litter depend very much on

prevailing environmental conditions. For example, moisture conditions influence plant

responses to litter. In mesic environments litter usually reduces seedling establishment,

probably through light limitation (Facelli and Kerrigan 1996). As precipitation decreases,

113

reduced desiccation associated with litter becomes more important and, when water is

limiting, the more mesic conditions associated with litter may result in enhanced seedling

establishment (Fowler 1986; Whisenant 1990b). Xiong and Nilsson (1999) concluded from

their meta-analysis of litter studies that litter generally had a negative effect on vegetation.

However, the reported effects of litter on seedling germination and emergence vary

considerably depending on the species involved, litter characteristics, environmental

conditions, and study methods (see reviews by Facelli and Picket 1991a and Xiong and

Nilsson 1999).

The amount and distribution of litter can be manipulated by fire. Generally fire reduces litter,

although some fires increase litter where they do not consume the canopy but are hot enough

to kill the leaves (Facelli and Kerrigan 1996). Litter may be important in determining

seedling emergence patterns (Young et al. 1981) and by altering litter cover, fire may

significantly affect establishment patterns and species composition. To help evaluate fire as a

tool for manipulating the composition of C. ciliaris-dominated grassland, I investigated the

effects of litter on grass seed germination and seedling emergence. Of specific interest were

the litter-grass species interactions at the C. ciliaris dominated site at Dalrymple (section

3.3). At this site, soil cover was variable in terms of quantity as well as type, with conditions

ranging from no cover (bare soil) to complete litter cover. Litter consisted of dead

herbaceous material as well as leaf and other matter from trees and shrubs (predominantly

eucalypts) and these litter types were patchily distributed over the site. In some areas, a thick

litter mat formed between C. ciliaris tussocks. This litter consisted of C. ciliaris stem and

leaf material, some still attached to the plant, which formed a consolidated mat (Figure 3.7).

It contrasted with herbaceous litter in other patches that consisted of individual pieces of

stem and leaf material that had not become fused. Matted litter, which is held together by

fungal hyphae, is a greater impediment than loose litter (Facelli and Pickett 1991a). Thick

layers of litter have been found to hinder germination and/or emergence (Scanlan and

O’Rourke 1982; Fowler 1986; Hamrick and Lee 1987; López-Barrera and González-

Espinosa 2001) and have been associated with reduced plant productivity (Harrington and

Ross 1974; Knapp and Seastedt 1986). I expected that the matted C. ciliaris litter would

hinder or inhibit seedling emergence by preventing seedlings from emerging from beneath it

and/or by preventing the roots of surface-germinated seedlings from reaching the soil

surface.

The aims of this study were to determine (1) if soil surface cover affected the germination

and emergence of two grass species and (2) whether these effects differed between the two

species. The effects of four soil surface cover types on the germination and emergence of

114

two perennial grasses, C. ciliaris and Heteropogon contortus, were investigated. These are

key herbaceous species in this habitat. The exotic C. ciliaris is dominant in terms of biomass.

The native H. contortus is common at Dalrymple and is a desirable replacement for C.

ciliaris. Therefore, the interactions of these species with soil surface condition were of

interest. The diaspores, or dispersal units, of these two species are very different. Cenchrus

ciliaris has a burr type diaspore which initially prevents direct seed-soil contact. In contrast,

the diaspore of H. contortus has a hydroscopically active awn and sharply pointed base that

facilitate seed burial (Tothill 1969). The ability of H. contortus to bury into soil appeared

likely to give this species advantages in coping with litter. Seed material of both species

were sown under and on top of the different litter types to investigate how litter effects seed

dispersed prior to and after litter deposition. Sowing position was expected to have less

effect on H. contortus than on C. ciliaris. However, the observed litter effects and species–

sowing position interactions were contrary to expectations. Enhanced emergence from

matted litter, species differences in sowing position effects, and the potential effects of fire

on establishment patterns are discussed.

Figure 3.7. Litter mat between C. ciliaris tussocks at Dalrymple.

115

3.6.2. METHODS

3.6.2.1. Treatments

Seeds of two perennial grasses were sown on bare soil and under or on top of one of three

litter types. The treatments were: two species (C. ciliaris and H. contortus), four soil surface

cover types (eucalypt litter, open C. ciliaris litter, matted C. ciliaris litter and bare (no litter))

which represented the main surface cover types at the Dalrymple experimental site, and two

sowing positions (under and on top of litter).

3.6.2.2. Experimental details

The experiment was conducted on outdoor benches at Davies Laboratory, Townsville

(19o15’S, 146o45’E) in January and February 2001. Thirty-two free-draining trays (43 cm by

28 cm by 13 cm deep) were partly filled to a depth of about 7 cm with a loam-organic matter

mixture: three parts loam and one part “Aussie peat” (see appendix 2A). Trays were arranged

eight to a bench with each bench considered a block. Trays were watered to germinate any

pre-existing seed in the potting mix.

Litter was collected from areas beside the experimental plots at Dalrymple (section 3.3) in

December 2000. Three types of litter were collected: (1) eucalypt litter consisting of leaves

and twigs, (2) open grass litter consisting of C. ciliaris leaf and stem material and (3) matted

grass litter, consisting of consolidated C. ciliaris leaf and stem material forming litter mats.

The eucalypt litter was bagged while the matted litter was carefully removed from the soil

surface in ‘sheets’ and stored to ensure its structure remained intact. Stem and leaf material

was cut from C. ciliaris plants to make the open litter. The litter was wetted and then heated

in plastic bags at 60 oC for five days to kill any resident seed.

Cover types were randomly allocated to trays within benches (blocks), with two trays/bench

receiving each cover type. Eucalypt and open litter were spread evenly over the soil surface

and the matted litter sheets were cut to fit the trays (Figure 3.8). Litter in each tray was

allowed to dry before being adjusted to equal weight. (The mean weight of litter per tray was

1137 ± 29.7 g/m2. This fell within the range found for matted litter in the field (310-1385

g/m2 matted litter for 30 20 cm by 20 cm samples collected in January 2001 from the site).

Diaspores of H. contortus and C. ciliaris were sown in the trays on 31 January 2001. The H.

116

contortus seed had been collected from near the Dalrymple site in June 2000 while the C.

ciliaris was commercial seed (‘USA’ variety). Diaspores, rather than clean seed, were sown

to imitate field conditions. The diaspore of C. ciliaris is a fascicle comprising clusters of

spikelets (usually 1-3) surrounded by an involucre of two rows of wavy bristles (Loch 1993).

This burr type diaspore may contain one to five caryopses (Humphries 1981). The diaspore

of H. contortus is a fertile spikelet with a sharply pointed, callused base. Each fertile spikelet

contains two florets: a lower sterile floret (reduced to an empty lemma) and an upper fertile

floret, the lemma of which bears a stout, twisted and bent awn covered with short bristles in

the lower part (Tothill and Hacker 1983). The awned spikelets form tangled clumps of seed.

Seed material was sown, one species per tray of each cover type per bench. In the litter

treatments, material was sown on top of the litter at one end and under the litter at the other,

leaving an unsown buffer of 2.5 cm around the edge of the tray and a 6 cm buffer between

the two sowing positions. In bare soil trays, material was sown leaving the same buffer areas,

forming two sown plots per tray as in the other treatments. Species were allocated randomly

to trays within cover types. Sowing position was randomly allocated within trays. The

amounts of seed material sown (2.3 g of H. contortus ‘clumps’ or 0.71 g of C. ciliaris

fascicles) was estimated to provide about 40 germinable seeds per sowing position (see

appendix 2B), although these estimates are imprecise, particularly in the case of H.

contortus, because of the nature of the seed material. Trays were watered immediately after

the addition of seed material and were watered over the experimental period (rainfall and

hand watering with ‘rose’ head sprinkler).

Numbers of emerging seedlings were monitored for three weeks after sowing. Emerged

seedlings were counted every day for the first 10 days after sowing and then again 21 days

after sowing. Once seedling numbers exceeded about 30 it became difficult to make accurate

counts without disrupting the litter or seedlings. Therefore 30 or more seedlings were

recorded as 75% emergence. Germination and emergence were not investigated

independently and the observed responses are referred to as emergence responses for brevity.

In addition to the soil cover experiment, a survey of soil surface cover types at the Dalrymple

experimental site was conducted in January 2001. Thirty 20 cm by 20 cm quadrats were

surveyed in each control (unburnt) plot in the main experimental area. Quadrats were

positioned about every 3 m along five 14 m long transects spaced 4 m apart within each plot.

The percentage of each of the main cover types (bare soil, open, matted, and eucalypt litter)

in each quadrat was estimated to the nearest 25%.

117

Figure 3.8. Soil surface cover types with H. contortus seed. From left to right: eucalypt

litter, matted litter, bare soil and open litter.


Seedling emergence data were analysed using the repeated measures ANOVA procedure in

Genstat (2001). The time main effect and interactions were significant and more detailed

analyses were conducted on data for each sampling day separately. Data for each sampling

day were analysed using a modified split plot design, also using Genstat (2001). Species and

cover types were main plots. The cover type effect was partitioned to allow comparisons

between litter cover and no litter cover (the litter effect) as well as comparisons between

cover types. Sowing position was the subplot factor. However, there was no sowing position

effect for the bare soil treatment. This was accommodated in the analysis by partitioning

variances using litter (present/absent) by subplot interactions. Seedling counts were

converted to percentage emergence and arcsin transformed prior to analysis. Means were

compared using the protected LSD test at the 5% significance level. Reported results are

back-transformed means.

118

3.6.3. RESULTS

Litter had a significant effect on seedling emergence. Overall, emergence was greater from

bare soil than from litter (P < 0.001). However, there were significant differences between

cover types (Figure 3.9). Generally, emergence from bare soil was similar to that from

matted litter and emergence from these two cover types was greater than that from open and

eucalypt litter (P < 0.05). Initially emergence was greater from open litter than from eucalypt

litter but after day eight emergence from these two litter types was similar.

0

10

20

30

40

50

60

70

)

Per

cent

age

seed

ling

emer

genc

e (%

)

Figure

seedlin

litter (

The e

Emerg

sown

emerg

day 21

on see

The ra

experi

litter f

open l

top of

(a)

0

10

20

30

40

50

60

70

0 3 6 9 12 15 18 21

Days afte

3.9. Mean percentage emergence of (a)

gs from four cover types (bare soil (♦), ma

□)) over 21 days.

ffect of sowing position differed between

ence from seed sown on top of eucalypt li

under eucalypt litter. The reverse was fou

ence compared with under-litter sowing, alt

there was no sowing position effect for m

dling emergence from open litter over the ex

nking of sowing treatments for seedling e

mental period (Figure 3.10). Emergence wa

ollowed by seed sown on bare soil, under

itter, on top of open litter and on top of euc

matted litter was consistently greater than th

(b

0 3 6 9 12 15 18 21

r sowing

C. ciliaris seedlings and (b) H. contortus

tted litter (∆), open litter (X) and eucalypt

cover types (P < 0.05) (Figure 3.10).

tter was consistently lower than from seed

nd for matted litter: top sowing enhanced

hough initially (days three and four) and at

atted litter. Sowing position had no effect

periment (P > 0.05).

mergence was generally constant over the

s greatest from seed sown on top of matted

matted litter, under eucalypt litter, under

alypt litter. Emergence from seed sown on

at from seed sown on top of or under open

119

and eucalypt litter (P < 0.05). Emergence was similar from seed sown under eucalypt litter,

under open litter and on top of open litter while emergence from seed sown on top of

eucalypt litter was lower than from other treatments (P < 0.05) except on day 21 where it

was similar to that for seed sown on top of open litter.

0

10

20

30

40

50

60

70

0 3 6 9 12 15 18 21

Days after sowing

Per

cent

age

seed

ling

emer

genc

e (%

)

Figure 3.10. Mean percentage emergence of C. ciliaris plus H. contortus seedlings from four

cover types (bare soil (♦), matted litter (∆), open litter (X) and eucalypt litter (□)) from seed

sown top of litter (solid line) and under litter (dashed line) over 21 days.

There were differences in emergence patterns between species. More H. contortus than C.

ciliaris seedlings were found on all sampling days (P < 0.05). While litter depressed the

emergence of both species it appeared to have a greater effect on C. ciliaris. Significant

species by litter interactions were detected initially (day three) and later (days nine and 10)

(P < 0.05) where emergence from litter compared with bare soil was about 66% for C.

ciliaris and 88% for H. contortus. Species responded differently to sowing position with

significant species by sowing position interactions found on all survey days (P < 0.05)

(Figure 3.11). Initially sowing position had no effect on C. ciliaris emergence but after day

five more C. ciliaris seedlings emerged from seed sown on top of litter than from seed sown

under litter. By day 21 there was no difference in C. ciliaris emergence between sowing

positions. Heteropogon contortus emergence was affected by sowing position throughout the

experiment but, in contrast to C. ciliaris, emergence was greater from seed sown under litter.

120

Although species by litter type by sowing position effects were not significant (P > 0.05), C.

ciliaris tended to be more responsive than H. contortus to sowing position with matted litter.

The reverse trend was found for eucalypt litter, that is H. contortus tended to be more

responsive than C. ciliaris to sowing position with eucalypt litter.

0

10

20

30

40

50

60

70

0 3 6 9 12 15 18 21

Days after sowing

Per

cent

age

seed

ling

emer

genc

e (%

)

Figure 3.11. Mean percentage emergence of C. ciliaris seedlings (X), and H. contortus

seedlings (□) from seed sown on top of litter (solid line) and under litter (dashed line) over

21 days.

Emergence rates differed between treatments (Figure 3.12). Emergence was faster from bare

soil than from litter, taking on average 5.6 days for 50% of seeds to emerge. In comparison

emergence from some litter treatments did not reach 50% over the experimental period. H.

contortus emerged earlier that C. ciliaris. Two days after sowing, seedlings were found in

40% of H. contortus plots while no C. ciliaris seedlings had emerged. In the bare soil

treatment 50% seedling emergence took 3 days for H. contortus and 8 days for C. ciliaris.

Heteropogon contortus achieved 50% emergence in all treatment plots except where seed

was sown on top of eucalypt and open litter. In contrast C. ciliaris achieved 50% emergence

in only two treatments: bare soil and from seed sown on top of matted litter.

Although the estimated number of germinable seeds added to each plot was imprecise due to

the nature of the seed material, emergence from treatment replicates was generally very

similar. This may indicate that plots received a similar number of germinable seeds.

121

0 5 10 15 20

Eucalypt-top

Eucalypt-under

Open-top

Open-under

Matted-top

Matted-under

Bare soil

Days after sowing

Figure 3.12. Mean days taken for 50% emergence of C. ciliaris seedlings (lower solid bars)

and H. contortus seedlings (upper unfilled bars) from seed sown on bare soil and under and

on top of three litter types. Bars running the length of the graph indicate that 50% emergence

was not achieved in all replicates by day 21.

The distribution of litter cover types in the field was variable (Figure 3.13). For example, in

one plot no bare ground was recorded while in another, bare ground made up 33% of the

surveyed area.

0

10

20

30

40

50

Bare Eucalypt Open Matted

Soil cover type

Per

cent

age

soil

cove

r typ

e (%

)

Figure 3.13. Mean (± SE) percentage of soil cover types at Dalrymple in January 2001.

122

3.6.4. DISCUSSION

Litter type is important in determining the effects of plant litter (Xiong and Nilsson 1999)

and differences in litter structure have been shown to differentially affect seed germination

and/or emergence patterns (Hamrick and Lee 1987; Facelli and Pickett 1991b; Facelli and

Kerrigan 1996; López-Barrera and González-Espinosa 2001). Although the overall effect of

litter on emergence in this study was negative, as found by many others (Fowler 1988;

Bergelson 1990; Facelli and Pickett 1991b; Barritt and Facelli 2001), emergence of both C.

ciliaris and H. contortus was affected by litter type.

The enhanced emergence from matted litter compared with the other litter types was contrary

to expectations. Litter may form a physical barrier hindering seed movement and seedling

growth. This may be particularly so in the case of matted litter since it appears to present a

greater impediment than open litter (Facelli and Pickett 1991a). Seeds retained in litter may

have delayed or unsuccessful germination (Facelli and Pickett 1991a) and thick litter layers

have been reported to hinder the germination and/or emergence of C. ciliaris (McIvor and

Gardener 1981) and H. contortus (Kennard and Walker 1973) as well as other species

(Scanlan and O’Rourke 1982; Hamrick and Lee 1987). Dense litter slowed the downward

movement of Bouteloua rigidiseta and Aristida longiseta seeds to the soil surface, delaying

germination, and in the case of A. longiseta, reducing survival since some seed germinated in

the litter and died (Fowler 1986). However, it appears that the matted C. ciliaris litter used in

this study provided favourable conditions for germination and/or emergence and effects on

seeds and/or seedlings were less important. In contrast, the open and eucalypt litter appeared

to hinder seed movement and/or seedling growth, resulting in reduced emergence.

Seed-soil contact is often cited as a factor influencing germination (Cook et al. 1993b) but it

is actually the moisture conditions around the seed that are important. As noted by Harper et

al. (1965), differences in micro-topography exert effects on germination through modifying

seed-water relationships. It is likely that differences in seed-moisture conditions between

litter types explain the emergence patterns found in this study. The matted litter provided

suitable moisture conditions for germination. Its uneven but continuous surface enabled good

seed-litter contact and maintained moist conditions around the seeds. In addition, it dried out

more slowly than did bare soil or the other litter types, thereby maintaining moist conditions

for longer. Seed on the surface of matted litter germinated readily without needing contact

with the soil surface. In contrast, it appears that open and eucalypt litter surfaces did not

provide such conditions and most seed sown on top of these litter types had to reach the soil

surface for successful germination. While not forming a continuous layer, open and eucalypt

123

litter were deeper than the matted litter. Seeds sown on top of open and eucalypt litter had

less contact with litter surfaces, were exposed to less humid conditions and were hindered in

their movement to the soil surface. Constraints to seed movement and seedling growth are

the most likely cause of depressed seedling emergence from open and eucalypt litter.

However, since chemical and physical effects of litter were not differentiated in this study,

the possibility of chemical differences associated with the different litter types cannot be

dismissed. Some eucalypts, for example Eucalyptus globulus, have been shown to have

allelopathic effects (Babu and Kandasamy 1997).

Whereas emergence from seed sown under matted litter was generally greater than that from

other litter types it was less than that from top-sown seed, particularly for C. ciliaris. It is

unlikely that this was caused by light conditions under matted litter since neither C. ciliaris

(Chaudhry et al. 1999; Sharif-Zaden and Murdoch 2000) nor H. contortus (Orr 1998) have a

light requirement for germination. Conditions under matted litter would have been at least as

moist as conditions on top. Therefore desiccation would have been even less likely for seeds

sown under matted litter. The humid conditions associated with litter may promote disease

(Facelli et al. 1999) and emergence from under matted litter may have been reduced by

increased pathogen activity. An alternative or additional explanation is that matted litter

physically impeded seedling emergence.

Emergence from seed sown on top of eucalypt litter was lower than that from other

treatments. Litter from trees is believed to have a stronger effect on plants than grass litter,

although whether this relates to differences in chemical and/or physical properties is unclear

(Xiong and Nilsson 1999). Facelli and Kerrigan (1996), who also reported poor emergence

from seed sown on top of eucalypt litter, pointed out that eucalypt litter presents many

extended flat surfaces that retain seed. Seed falling on eucalypt litter may be trapped in

conditions unsuitable for germination while seed germinating in such litter may die if the

radicle is prevented from reaching the soil surface.

Open and eucalypt litter may also hinder seedling emergence from the soil surface.

Emergence was generally lower from seed sown under eucalypt or open litter than on bare

soil. Seed under litter was probably exposed to less desiccation than seed on bare soil so it is

unlikely that germination and/or emergence were constrained by moisture. While both litter

types were loose in structure, seedlings were forced to bend around obstacles before

emerging. Such energy-consuming emergence may weaken seedlings, resulting in increased

mortality (Hamrick and Lee 1987; Facelli and Kerrigan 1996). There may have been greater

mortality of seedlings emerging from under eucalypt and open litter. Since the fate of

124

individual seedlings was not followed in this study, it is not possible to accurately determine

mortality levels. Dead seedlings were observed and cumulative emergence decreased in

some plots. However, there were no obvious treatment effects on mortality as far as it could

be assessed. Facelli et al. (1999) reported that eucalypt litter did not impede the emergence

of eucalypt seedlings and an alternative or additional explanation for poor emergence from

under litter is that seeds under litter were exposed to greater disease.

Differences in litter structure have been found to affect germination rate (López-Barrera and

González-Espinosa 2001). In this study, seedling emergence was faster from bare soil and

matted litter than from open and eucalypt litter. Differences in germination rate may have

important consequences for plant establishment and the timing of germination is recognized

as an important factor influencing seedling survival and community composition (Fowler

1986; 1988). Early germination is usually considered to give an individual a competitive

advantage, although too early germination can be fatal in semi-arid and unpredictable

environments (Fowler 1986).

In addition to litter type, seed and seedling characteristics also influence plant responses to

litter. The size and shape of seeds influence seed movement through litter while seedling

morphology may be important in determining the ability of seedlings to penetrate litter

(Gross and Werner 1982; Facelli and Pickett 1991a). Although differences in emergence due

to C. ciliaris and H. contortus seedling morphology were not expected to be significant in

this study, differences in emergence due to diaspore type were. I expected sowing position to

have less effect on H. contortus than on C. ciliaris since H. contortus seed sown on top of

litter would work its way through to the soil surface, its movement being assisted by its

hydroscopically active awn and pointed seed base. In contrast, the burr type C. ciliaris

diaspore appeared more likely to be caught in litter and prevented from reaching the soil

surface. However, the observed sowing position effects were not as anticipated.

Heteropogon contortus emergence was affected by sowing position, emergence being greater

for seed sown under litter than on top. It appears that litter hindered the movement of H.

contortus seed, preventing at least some from reaching the soil surface. This result is

interesting given that H. contortus is believed to have an efficient seed burial mechanism

(Tothill 1969; Peart 1979; Dyer et al. 1997) which has been cited as giving it an advantage

over other species whose seed remain caught in litter (Tothill 1969). While H. contortus may

be able to penetrate litter more effectively than some other species, the results from this

study suggest that litter can have a negative impact on its emergence. Cenchrus ciliaris

emergence was also affected by sowing position, but its emergence was greater for seed

sown on top of litter. This effect tended to be strongest for matted litter, although species by

125

litter type by sowing position effects were not statistically significant. Poor seed-soil contact

has been considered a problem in C. ciliaris establishment (Cook 1975) and emergence was

expected to be greater for seed sown under litter. Good germination and establishment of C.

ciliaris requires seed to be moist for about five days and a rough surface is considered

advantageous for C. ciliaris establishment since it maintains moisture for longer (Paull and

Lee 1978). The surface of the matted litter provided favourable germination conditions and

seedlings were able to penetrate the litter to reach the soil. In contrast, reduced emergence of

C. ciliaris from seed sown under litter may reflect greater exposure to disease and/or

impeded seedling emergence. Hacker (1989) suggested that greater levels of insect and

microbial activity associated with litter could lead to deterioration of C. ciliaris seed in the

field.

The markedly faster emergence of H. contortus compared with. C. ciliaris may have

important implications for plant establishment patterns. Whereas C. ciliaris is considered a

highly competitive species, it is less so at the seedling stage when seedling survival is

strongly affected by competition from established plants (Cook 1984; McIvor, 2003). The

slower emergence of C. ciliaris compared with H. contortus suggests that C. ciliaris may

also be at a competitive disadvantage against other seedlings. The faster emergence of H.

contortus may give it a head start in sequestering resources and enable it to out-compete C.

ciliaris for the occupancy of new sites.

This study was conducted to help predict the effects of fire on seedling emergence patterns.

Although the results demonstrate that the removal of litter by fire is likely to effect

emergence patterns in this C. ciliaris-dominated grassland, it is difficult to predict the

outcome of litter removal. Extrapolation to the field of seedling emergence-litter interactions

determined under glasshouse conditions is complicated by the fact that environmental

conditions can greatly influence plant responses to litter. Glasshouse studies in which

moisture conditions are maintained may under-estimate the benefits of litter in arid and semi-

arid field conditions (Fowler 1986). As moisture becomes limiting the importance of litter

for reducing desiccation increases (Facelli and Kerrigan 1996) and litter has generally been

found to have positive effects on emergence in arid and semi-arid environments (Facelli and

Kerrigan 1996). McIvor and Gardener (1985) reported higher C. ciliaris emergence in

unburnt compared with burnt pastures. This was associated with the beneficial effects of

cover in the vegetated plots that resulted in longer germination periods. When moisture is

limiting to growth, matted-litter sites may become significantly more favourable for seedling

emergence than bare ground. The deep layers of open and eucalypt litter used in this study

inhibited seedling emergence. However, litter distribution in the field varies greatly and it is

126

possible that lighter levels of these litter types may also have positive effects on emergence

in the field. Plant responses to litter have been found to vary with changes in the amount of

litter (Scanlan and O’Rourke 1982). Therefore, while overall, litter was found to depress

seedling emergence, suggesting that removal of litter by fire will result in increased

emergence, the effects of litter removal depend on nature and distribution of litter as well as

the prevailing conditions. An additional factor to consider is the direct effect fire has on seed

availability. Emergence may be greater from litter microsites because litter traps seed,

increasing its availability in these sites (Scanlan and O’Rourke 1982). Both C. ciliaris

(Hacker and Ratcliff 1989) and H. contortus (Tothill 1977) have dormancy mechanisms

preventing the germination of freshly fallen seed so much seed is likely to be caught in and

under litter by the time it is ready to germinate. Therefore, fire is likely to affect emergence

patterns by reducing seed availability (McIvor et al. 1993).

Litter, via its effects on establishment patterns, may play an important role in population

dynamics (Bergelson 1990), inter-specific interactions (Al-Mufti et al. 1977, Facelli 1994)

and community structure and composition (Knapp and Seastedt 1986; Hobbs and Aitkins

1988; Facelli and Pickett 1991ab). By altering the distribution and abundance of litter, fire

has been attributed with affecting grassland productivity and species richness (Knapp and

Seastedt 1986; Xiong and Nilsson 1999). The prevailing fire regime may favour some

species over others via litter effects on plant establishment. For example, Bromus spp.

establishment in American prairies has been found to be enhanced by litter (Evans and

Young 1970; Whisenant 1990b) and increases in Bromus abundance have been associated

with low fire frequency which maintains a favourable litter status. In contrast, emergence of

H. contortus in southern Queensland is believed to be promoted by the removal of litter and

consequent rise in soil temperatures (Tothill 1969). Of particular interest in this study was

whether C. ciliaris and H. contortus emergence would be differentially affected by litter

removal. This remains unclear. By removing litter fire may increase C. ciliaris emergence

since overall litter had a greater negative effect on C. ciliaris than on H. contortus. However,

this effect was not marked or consistent and other observed emergence patterns suggest that

C. ciliaris may be disadvantaged by litter removal. Cenchrus ciliaris seeds are likely to be

present in matted-litter sites since this litter type forms around established C. ciliaris plants.

Consequently, the removal of matted litter is likely to have a greater negative effect on C.

ciliaris emergence than on H. contortus emergence. The consumption by fire of seed in litter

may affect both species. While such loss was expected in the case of C. ciliaris, the finding

that H. contortus seed may be trapped in litter suggests that this species may be more

vulnerable to seed loss via fire than previously thought.

127

Understanding how species, litter, precipitation and fire interact is necessary for developing

fire management strategies (Whisenant 1990b). Although it is likely that litter affects the

early establishment of C. ciliaris and H. contortus, the responses to litter removal by fire

depend on the nature and distribution of litter as well as the prevailing conditions. From this

study it is difficult to identify any major, differential effects of litter removal on the

emergence of C. ciliaris and H. contortus. In the field, litter may improve seedling

emergence due to the combination of improved moisture conditions, lower soil temperatures

and the greater quantity of germinable seed where there is more litter (Scanlan and O’Rourke

1982). However, sites favourable for seedling emergence may not necessarily be favourable

for plant establishment (Schupp 1995) since, once established, other factors become

important (Gross and Werner 1982). Whereas matted litter may favour seedling emergence,

seedling survival and establishment in matted litter sites may be highly unlikely given the

competition from surrounding, established plants. In perennial grasslands, the opportunity

for new individuals to establish is greatly influenced by competition from established plants

(Cook 1980; Lauenroth and Aguilera 1998). Therefore, while fire may affect emergence

patterns via its effects on litter, fire effects on established plant survival are likely to play a

more significant part in determining vegetation composition.

128

3.7. EFFECTS OF SEASON OF BURNING ON SOIL SEED BANKS OF CENCHRUS

CILIARIS-DOMINATED GRASSLAND

3.7.1. INTRODUCTION

The soil seed bank is the reserve of viable seed in the soil and on the soil surface (Roberts

1981). The study of soil seed banks has become an integral part of plant ecology (Grime

1989) and a knowledge of seed bank dynamics is imperative for predicting ecosystem

responses to management and natural perturbations (Roberts 1981). Fire may alter both the

size and composition of soil seed banks and these effects may result in changes in vegetation

composition since the seed bank determines the potential size and composition of the

seedling population (McIvor and Gardener 1991). In the context of using fire as a tool to

manipulate vegetation composition, an understanding of both the nature of seed banks and

their responses to fire is important for predicting vegetation change.

Fire affects soil seed banks directly, by killing seed and altering seed germinability (section

3.2.2.5). Although it may consume or damage seed, and many studies report reductions in

seed banks after fire (’t Mannetje et al. 1983; DiTomaso et al. 1999; Ferrandis et al. 1999;

Holl et al. 2000; Main et al. 2000; Alexander and D’Antonio 2003), it may increase the

amount of germinable seed (Purdie and Slatyer 1976). Many species require a heat shock to

permit germination (Bell 1999) and other factors, such as smoke, may also stimulate

germination (Enright et al. 1997; Roche et al. 1998; Read et al. 2000). Fire affects seed

banks indirectly, by affecting flowering and seed inputs (sections 3.2.2.5 and 3.8). Fire-

stimulated flowering is very common in grasses (Bond and van Wilgen 1996) and increases

in post-fire grass seed production have been reported (Orr 1998). The effects of fire on the

activity of microbes and animals may also have consequences for seed banks. For example,

litter has been associated with greater levels of microbial and insect activity (Hacker 1989;

Watson 1995) and its removal by fire may reduce seed losses due to disease and predation.

Since fire may affect both the quantity and quality of seed, it may simultaneously decrease

the total amount of seed but increase the amount of germinable seed present.

As well as affecting seed bank size, fire may alter seed bank composition where species

differ in their response to fire (Auld and Bradstock 1996). Characteristics such as seed

morphology (Peart 1984), heat tolerance, germination requirements and the location of seed

in the soil profile (refer Tyler 1995; Noble and Grice 2002) will influence how species are

affected by fire. For example, differences in seed burial capacity make species more or less

129

susceptible to fire. Since the lethal temperatures generated by grassland fires are generally

confined to the upper soil layer and the soil surface (Morgan 1999), species whose seed

remain on the soil surface will be more vulnerable than species whose seed is buried.

Burning is expected to reduce the abundance of C. ciliaris seed since it appears to remain

close to or on the soil surface (Hacker 1989), making it more vulnerable to fire than species

such as Heteropogon contortus that have effective seed burial mechanisms such as

hygroscopically active awns (Walker et al. 1981), although awned diaspores may not always

achieve sufficient burial (Peart 1984). Heat tolerance also varies between species. Many

Australian legume species are extremely tolerant of high-intensity or long-duration thermal

stress (Bell 1999) enabling them to survive fires that destroy less tolerant species. Therefore,

in a single fire, seeds of some species are destroyed while the germination of others is

promoted (Purdie and Slater 1976).

Interactions between fire characteristics and seed bank characteristics determine the effects

of fire on soil seed banks. For example, Auld (1986) described how the fate of Acacia

suaveolens seed depended on the interactions between seed depth in the soil and fire

intensity and duration. Fire characteristics such as temperature and duration are major factors

determining the outcome of burning on soil seed banks. Morgan (1999) concluded that fires

in temperate Australian grasslands probably killed most seed on the soil surface since lethal

surface temperatures were achieved. In contrast, fires in Mitchell grasslands in north-western

Queensland were unlikely to be detrimental to buried seed or seed on the soil surface that

were not directly consumed by fire since these fires exhibited low peak temperatures of short

duration (Scanlan and O’Rourke 1982). Season of burning may also significantly affect seed

banks. For example, in open woodland in New South Wales, soil temperatures are higher

after summer fires than after winter fires, influencing the numbers of germinable seeds in the

seed bank (Auld and Bradstock 1996).

The aim of this study was to investigate the effects of season of burning on the germinable

soil seed banks of C. ciliaris-dominated grassland in two vegetation communities in north

Queensland. Seed banks may play a significant role in determining vegetation composition,

especially following disturbance (Warr et al. 1993) and an understanding of the nature of

these seed banks and their responses to fire will help predict how the composition of these

communities may be affected by fire.

Methods used to determine the numbers and identity of seeds in seed banks can be divided

into two groups, germination methods and physical extraction methods, each with their

strengths and limitations (Roberts 1981; Simpson et al. 1989). The identification of species

130

is generally easier using the germination method. However, it is only effective for

determining the germinable seed bank. In contrast, physical extraction methods enable

determination of viable seeds. A further limitation of the germination method is it may

under-estimate seed numbers of some species since the specific germination requirements of

all species present are unlikely to be met by a particular germination technique (Roberts

1981). Despite these limitations, the germination method is suitable for this study and was

the method used since, as pointed out by Clifford et al. (1995), it is the immediately

germinable flora that determines the response to disturbance.

3.7.2. METHODS

3.7.2.1. Soil collection and processing

Soil was collected from the experimental plots at Dalrymple and Moorrinya after the late dry

season fires in November 1999 (section 3.3). Samples were collected at this time to

determine the seed banks present at the start of the growing season. At Dalrymple, four soil

cores (2.8 cm diameter, 5 cm depth) were collected at the corners of a 40 cm by 40 cm

quadrat at each of 12 randomly located sampling positions within each plot. A different soil

collection method was used at Moorrinya since the soil corer used at Dalrymple was not

suitable for sampling the Moorrinya clay. At Moorrinya one soil sample (7.2 cm diameter, 5

cm depth) was collected from each of eight randomly located sampling positions by

hammering an open-ended metal cylinder into the soil until it was flush with the soil surface.

The cylinder was dug up and the soil removed. While every care was taken, it was difficult

to collect consistently sized samples because of the aggregated nature of the cracking clay.

Samples collected within a plot were bulked. The bulked plot samples represented 296 cm2

soil surface per plot for Dalrymple and 326 cm2 soil surface per plot for Moorrinya.

Each soil sample was sieved through a 6.7 mm sieve before further processing.

3.7.2.2. Glasshouse set up

The composition of the soil seed banks was determined using a germination method. Eight

plastic pots (approximately 13 cm diameter, 13.5 cm high) per soil sample were set up in a

shade house at Davies Laboratory, Townsville (19o15’ S, 146o45’ E), in November 1999.

131

Pots were lined with shade cloth to prevent soil loss from the drainage holes and partially

filled with sterilized fine sand (autoclaved at 121oC for 20 minutes). For each soil sample,

the soil was spread over the sand of eight pots to form a layer about 2.5 cm thick per pot.

Pots were randomly allocated to water-tight trays to which water was added to maintain

moist conditions in the pots by capillary action over the experimental period. Trays were

kept in ambient light conditions and pots were re-randomized during the experiment.

3.7.2.3. Measurements

The identity and number of seedlings that emerged from each pot were recorded over an

eleven month period. Most seedlings were identified to species level. Seedlings were

removed once they could be identified and were dried to produce a herbarium collection and

assist with identification of unknown specimens. Self-sowing occurred in some plots since

seedlings were generally grown until they flowered to aid identification. Self-sown plants,

which emerged later than original individuals and generally in high densities, were excluded

from the analyses.


Burning treatment effects on total numbers of seedlings and numbers of annual grass,

perennial grass, legume, forb (referring to non-leguminous forbs) and sedge seedlings as

well as numbers of seedlings of some individual species (those for which sufficient data

were available) were compared using one-way ANOVA. Seedling counts per plot were

square-root transformed prior to analysis and the reported means have been back-

transformed and converted to counts/m2. Burning treatment effects on species richness

(number of species/400 m2 plot) and diversity (as measured by Simpson’s index) were also

compared using ANOVA. Simpson’s index is calculated as:

1

Where s is the total number of s

individuals of species i of the total i

plot fire intensity and species richne

number were investigated using

D =

ΣPi2

s i=1

pecies in the community and Pi is the proportion of

n the sample (Begon et al. 1990). Relationships between

ss and between plot fire intensity and seed bank seedling

linear regression. All analyses were performed using

132

Genstat (2001).

The proportional composition of the seed banks was compared graphically to the

proportional cover in the standing vegetation. Collection and analysis of the standing

vegetation data are described in section 3.10.

3.7.3. RESULTS

Burning appeared to have little effect on the germinable seed banks of either site. No burning

treatment effects on numbers of perennial grass, annual grass, forb, legume or sedge

seedlings (Figure 3.14) or total seedling numbers were detected (P > 0.05). Total seedling

numbers ranged from 271-846 (mean 480) seedlings/m2 for Dalrymple seed banks and 276-

1351 (mean 606) seedlings/m2 for Moorrinya seed banks.

Few C. ciliaris grass seedlings emerged (five and four seedlings from the Dalrymple and

Moorrinya seed banks respectively). Perennial grasses were a minor component of the seed

banks, making up 22% and 12% of the Dalrymple and Moorrinya seed banks respectively.

Interestingly, despite its dominance in the standing vegetation, no Astrebla seedlings

emerged from the Moorrinya seed banks. The seed banks of both sites were dominated by

forb seedlings which made up 53% and 41% of the Dalrymple and Moorrinya seed banks

respectively (Figure 3.15). This contrasts with the composition of the standing vegetation

which was dominated, in terms of cover, by perennial grasses (Figure 3.15). The 175

seedlings that germinated from the Dalrymple seed banks included three annual grass, nine

perennial grass, 25 forb, three legume and five sedge species. The 253 seedlings that

germinated from the Moorrinya seed banks included four annual grass, seven perennial

grass, 21 forb, two legume and four sedge species. (Identified species are listed in

appendices 3A and 3B).

Burning treatment effects could not be investigated for most species individually because of

low seedling numbers. About three quarters of the species identified were represented by less

than five individuals. Eleven species had 10 or more individuals and of these Iseilema spp.

were by far the most abundant with 66 seedlings, the next most abundant species having less

than 20 seedlings. Seedling emergence of these ‘abundant’ species was generally highly

clumped with a high proportion of seedlings emerging from single plots. For example, more

than half of the Iseilema seedlings that emerged from the Moorrinya seed banks were from

133

0

50

100

150

200

250

300

350

Control Early Late

0

50

100

150

200

250

300

350

Control Early Late

0

50

100

150

200

250

300

350

Control Early Late

0

50

100

150

200

250

300

350

Control Early Late

0

50

100

150

200

250

300

350

Control Early Late

0

50

100

150

200

250

300

350Dalrymple Moorrinya (a)

Control Early Late

0

50

100

150

200

250

300

350

ontrol Early Late

(b)

C

0

50

100

150

200

250

300

350

ontrol Early Late

(c)

C

0

50

100

150

200

250

300

350

ontrol Early Late

(d)

C

0

50

100

150

200

250

300

350

Control Early Late

Burning treatments

(e)

Mea

n nu

mbe

r of s

eedl

ings

(cou

nts/

m2 )

Figure 3.14. Mean number (counts/m2) of (a) annual grass, (b) perennial grass, (c) forb, (d)

legume and (e) sedge seedlings from unburnt (control), early dry season and late dry season

burnt seed banks from Dalrymple (1st column) and Moorrinya (2nd column).

134

one plot. Eight of the 14 Indigofera colutea seedlings, eight of the 17 Corcorus tridens

seedlings and nine of the 14 Eragrostis brownii seedlings also emerged from a single plot.

Only three of these abundant species occurred in more than half the plots from a site. The

highly clumped pattern of emergence made detecting treatment effects difficult and no

burning treatment effects on seedling counts were found for the species investigated (P >

0.05).

Dalrymple Moorrinya

0

20

40

60

80

100

Seed bankseedlings

Standingplant cover

Seed bankseedlings

Standingplant coverP

erce

ntag

e co

mpo

sitio

n of

see

d ba

nk s

eedl

ings

or

stan

ding

pla

nt c

over

(%)

Perennial grass Annual grass Forb Legume Sedge

Figure 3.15. Percentages of perennial grass, annual grass, forb, legume and sedge plant

groups making up the seed bank seedlings (1st bar) and post-fire standing herbaceous plant

cover (2nd bar) at Dalrymple (left) and Moorrinya (right).

Species richness and diversity (Simpson’s index) tended to be higher in the burnt treatments

compared with the unburnt treatment for the Dalrymple seed banks, although this effect was

not statistically significant (P = 0.053 and 0.08 for species richness and diversity

respectively) (Figure 3.16). Burning treatment effects on the seedling species richness and

diversity were not significant for the Moorrinya seed banks (P > 0.05) (Figure 3.16).

135

0

2

4

6

8

10

12

Control Early Late Control Early LateBurning treatments

D=6.0

D=6.8D=4.8D=6.2

D=6.8

D=3.5

Moorrinya Dalrymple

Num

ber o

f spe

cies

/plo

t

Figure 3.16. Mean species richness (number of species/plot) of seedlings emerging from

unburnt (control), early dry season and late dry season burnt seed banks from Dalrymple and

Moorrinya. Mean Simpson’s diversity index values (D) are also presented (SE = 0.95 and

1.22 for Dalrymple and Moorrinya diversity index values respectively).

There was a significant linear relationship (P < 0.05) between plot fire intensity and plant

species richness at Dalrymple (Figure 3.17). However, intensity explained only 36% of the

variation in species richness and the relationship was not significant when the high intensity

plot, which had high leverage, was omitted from the analysis. The intensity-species richness

relationship was not significant for the Moorrinya plots (P = 0.06) (Figure 3.17). No

relationships between fire intensity and number of seedlings were detected (P > 0.05).

0

2

4

6

8

10

12

14

16

0 1000 2000 3000

Num

ber o

f spe

cies

Richness=0.002*Fire Intensity + 6.78 Adj. R2 = 0.24

(b)

0

2

4

6

8

10

12

14

16

0 1000 2000 3000

Num

ber o

f spe

cies

/plo

t

Richness=0.002*Fire Intensity + 7.28 Adj. R2 = 0.36

(a)

Plot fire intensity (kW/m)

Figure 3.17. Relationships between plot fire intensity (kW/m) and plot species richness

(number of species/plot) for (a) Dalrymple and (b) Moorrinya seed banks.

136

3.7.4. DISCUSSION

The germinable seed banks of many communities are affected by fire (Shea et al. 1979; Mott

and Andrew 1985b; Ferrandis et al. 1999; Holl et al. 2000). However, the fires imposed in

this study appeared to have had little effect on the germinable seed banks of these C. ciliaris-

dominated communities. No burning treatment effects on total emergence or the emergence

of any plant group were detected. The seed bank densities found here fall within the range

reported for other communities in north-eastern Queensland (McIvor and Gardener 1991).

However, different methods of seed bank assessment make comparisons between studies

difficult (Warr et al. 1993).

Both increases (Calvo et al. 1999) and decreases (Morgan 1998; Holl et al. 2000) in seed

bank species richness with burning have been reported. In this study, burning may have

increased species richness at Dalrymple, although the results were inconclusive. Trends of

increasing species richness with increasing fire intensity were also inconclusive.

Fire reduces seed abundance by consuming or damaging seed. The apparent lack of response

to burning found here may indicate that much of the seed in these communities is protected

from fire. However, this seems unlikely for early dry season fires at least. Many herbaceous

species in northern Australia do not form persistent seed banks and much of the germinable

seed bank is produced over the previous growing season (McIvor and Gardener 1994).

Consequently, some loss of seed with early dry season burning is expected since the seed has

had relatively little time to become incorporated into the soil. For example, early dry season

fires are likely to burn H. contortus seed whilst in the seed heads or on the soil surface

(Walker et al. 1983). In contrast, late dry season fires will have less effect since 90% of H.

contortus seed becomes buried below the soil surface by spring (Campbell 1995). The

germinable seed banks of these C. ciliaris-dominated communities appeared unaffected by

season of burning. These results suggest that the fires did not reduce germinable seed

abundance. However, since thick litter, when present, was removed from the edges of

samples to enable soil coring, seed trapped in litter may have been under-represented in the

sampling procedure, and consequently, reductions in seed abundance were not detected.

Species differ in their response to fire and changes in seed bank composition with burning

were expected. Cenchrus ciliaris appears vulnerable to fire. Its diaspores have no effective

seed burial mechanism and seem to remain on or close to the soil surface (Hacker 1989).

Ernst (1991) concluded that C. ciliaris on the soil surface would be destroyed by fire due to

the high flammability of its glumes and appendices, whilst below the surface it is susceptible

137

to damage by dry heat. This contrasts with other species, such as Heteropogon contortus,

that have effective seed burial mechanisms (Walker et al. 1981) and may be stimulated to

germinate by dry heat (Zacharias et al. 1988). However, since few seedlings of these species

emerged, treatment effects could not be tested. There were fewer Iseilema seedlings in burnt

seed banks from Moorrinya, suggesting that Iseilema seed was killed by fire. However, the

highly clumped pattern of emergence may have prevented any statistical differences being

detected.

Fire also alters seed germinability, promoting the emergence of particular species and

groups, and many studies report that seeds of certain species germinate in greater quantities

after fire (Auld 1986; Shea et al. 1979; Purdie and Slater 1976). For example, Williams et al.

(2003b) reported increased emergence of the legumes Galactia tenuiflora and Indigofera

hirsuta and grasses as a group in coastal woodlands in north-eastern Queensland. No plant

group or species in the seed banks of these C. ciliaris-dominated grasslands appeared

responsive to burning. The number of legumes emerging from these seed banks was

surprisingly low given the expectation of fire-promoted legume germination. Although there

was a trend of more legume seedlings in burnt compared with unburnt seed banks from

Dalrymple, this was due to high emergence of Indigofera colutea in one plot and statistical

differences were not detected. Numbers of legume seeds have been found to be low

compared with those of other plant groups in a number of studies (refer Roberts 1981;

McIvor and Gardener 1994). However, others have reported that legumes may build up large

seed banks (refer Roberts 1981; Young et al. 1981). Reasons for the low numbers of legumes

found in these seed banks are unclear.

Although no major burning effects were detected, it is not possible to conclude that burning

had no effect on the germinable seed banks of these communities. There are sampling and

methodological issues that significantly hinder seed bank determination and these issues

were relevant for this study. Firstly, the variability in emergence patterns displayed by the

more abundant species greatly limited the power of the statistical tests. This clumped

distribution is typical of seed banks (Kemp 1989; Morgan 1998). Seeds are unevenly

distributed and this is recognized as a major problem for sampling (Bigwood and Inouye

1988; Warr et al. 1993). Large numbers of samples may be required to adequately describe

soil seed banks (Warr et al. 1993). Clearly, the sampling regime used here was inadequate to

account for the variability in these tropical grassland seed banks. However, it was not

feasible to increase the number of replicates (plots) in this study and it is unclear whether

more intensive within-plot sampling, which was precluded by time constraints, would have

increased the likelihood of detecting treatment effects. Secondly, seed numbers were

138

probably under-estimated by the germination method used since it is unlikely that optimal

conditions for the germination of all species present were achieved (Warr et al. 1993).

Comparison of seed bank determination methods have shown that different methods give

different results (Gross 1990). Emergence from pots watered by capillary action, as in this

study, has been found to be lower than from overhead-watered pots in some circumstances

(Orr et al. 1996), probably because overhead watering simulates field conditions more

closely by allowing fluctuations in moisture potential. Other studies have used multiple

germination runs with cycles of wetting and drying (e.g. McIvor and Gardener 1994). The

use of capillary watering and a single germination cycle may have resulted in under-

estimation of the germinable seed banks in this study.

An additional limitation of the germination technique is that it does not distinguish between

fire effects on seed abundance and seed germinability. Fire may simultaneously decrease

seed abundance but increase seed germinability resulting in similar numbers of seedlings

from burnt and unburnt seed banks (Mott 1982). Although it is unclear whether this

happened in this study, it would appear unlikely since significant differences in species

composition between burnt and unburnt seed banks would be expected if this had occurred.

Seed banks may be important in determining the potential vegetation composition following

disturbances such as fire. However, seed bank studies may be of limited value for predicting

vegetation change. A major problem with most seed bank studies is that estimates of seed

numbers are very imprecise (Bigwood and Inouye 1988). In addition to the problems posed

by spatial variability, seed banks may also vary considerably over time (McIvor and

Gardener 1994; Kemp 1989), particularly if they are strongly influenced by climate

variability and its effects on seed production and loss (Morgan 1998). Consequently, seed

banks are difficult to study from a sampling perspective and, as mentioned above, there are

also difficulties in species determination with different methods giving different results

(Roberts 1981; Gross 1990). Various recommendations proposed to better describe soil seed

banks include sampling over time rather than once only, collecting many small samples

rather than a few large samples and measuring seed production and/or seed rain (Roberts

1981; Thompson 1992). However, these suggestions are often impractical to implement. It is

also important to recognise that although seed bank studies may provide useful information

regarding the potential composition of seedling cohorts, the results may not reflect

emergence patterns observed in the field. For example, Shaw (1957) found that while fire

promoted H. contortus emergence in the field, more seedlings came up in the unburnt

treatment when the same soil was investigated in a glass house study. (The field effect was

probably due to increased soil temperatures after fire, see section 3.2.2.5). Despite the fact

139

that sampling and methodological issues limit the conclusions that can be made regarding the

effects of fire on these seed banks, it is likely that the patchiness and relatively low

intensities of the fires minimized any burning effects. The fires probably resulted in

relatively little destruction or damage to seed and had only minor stimulatory effects on

germination.

Although this study is inconclusive regarding the effects of season of burning on the

germinable seed banks of these two communities, it does provide useful data regarding the

soil seed banks themselves. Of particular interest is the apparent poor representation of

perennial grasses in the soil seed bank despite their abundance in the standing vegetation.

The composition of these seed banks reflects the commonly reported phenomenon of

germinable seed bank composition differing significantly from that of the standing

vegetation (McIvor and Gardener 1994; McIvor 1987; Morgan 1998; Roberts 1981; Rice

1989; refer Thompson 1992) and perennial grasses have been found to be poorly represented

in the seed banks of other perennial grass-dominated communities (eg McIvor and Gardener

1991, 1994; Lunt 1990; Everson 1999). Another consistent feature of grassland seed banks is

the presence of appreciable numbers of viable seeds of dicotyledonous plants (Roberts 1981;

Rice 1989). The seed banks of these C. ciliaris-dominated communities were also dominated

by forbs.

Differences in composition and richness of seed banks compared with the standing

vegetation could not be investigated in detail due to inconsistencies in plant identification

between this study and the study of vegetation composition (section 3.10). However, species

in the standing vegetation were not represented in the seed bank and vice versa. Species

richness was lower in the seed banks compared with the standing vegetation, as has been

reported elsewhere, for example, temperate Australian grasslands (Morgan 1998). The lack

of correspondence between the composition of the germinable seed bank and the standing

vegetation may be due to many factors including different levels of seed production, seed

mortality and predation, seed dissemination, dormancy characteristics and germination

requirements (McIvor 1987; McIvor and Gardener 1994). Differences in seed abundance

may reflect regeneration strategies in that species with high turnover in the vegetation are

more frequent in the seed bank than species with lower turnover. Most native perennial

species can sprout after fire and may offset the need for a large seed bank by long vegetative

persistence (refer Morgan 1998). In contrast, annual forbs and grasses which must regenerate

from seed each year are reliant on a soil seed bank for regeneration and generally have large

seed banks.

140

Despite the expectation of relatively low numbers of perennial species, the small numbers of

C. ciliaris seedlings to emerge from both communities and the absence of any Astrebla

seedlings from the Moorrinya seed banks was surprising and raises an important question:

why did so few perennial grass seedlings emerge? Either they are absent from the seed bank

or the sampling methodology failed to detect them. With regard to sampling methodology,

two issues are important. Firstly, assuming that the seed banks are spatially heterogeneous

and contain relatively few seeds of perennial species, it is likely that more perennials would

have been found with greater sampling effort. Secondly, it is possible that the germination

method used was unfavourable for the perennial species present. Samples were kept wet

continuously whereas cycles of wetting and drying may have been more favourable for

emergence of perennial grasses. For example, C. ciliaris is not tolerant of water-logging

(Anderson 1972) and the soils may have been kept too wet for C. ciliaris emergence.

Alternatively, the low numbers of perennial grass seedlings found may reflect actual low

abundance of germinable, perennial grass seed in these seed banks. Many of the dominant

grass species in northern (Mott and Andrew 1985b; McIvor and Gardener 1994) and

southern (Morgan 1998) Australian grasslands fail to form long-term persistent seed banks.

Cenchrus ciliaris seed loses most of its viability within two years under field conditions

(Silcock and Smith 1990), although one study reported seed remaining viable for up to four

years (Winkworth 1971). Astrebla also forms a transient seed bank because it lacks any long-

term dormancy mechanisms (refer Orr 1998). Since these species have no long-term storage,

seed numbers at the start of the wet season reflect the seed set the previous year and survival

over the dry season (McIvor and Gardener 1994). Consequently, the size of their seed banks

is largely dependent on the previous season’s growing conditions; low seed abundance may

reflect poorer than average growing conditions the previous season (Morgan 1998). Seed

banks have been found to be very variable from year to year (McIvor 1987). Seed input was

not measured in this study and its role in determining perennial grass seed banks cannot be

determined. However, rainfall was probably above average at both sites (nearby sites

experiences above average wet-season rainfall, section 3.3.) and many C. ciliaris and

Astrebla plants were in flower in 1999 (section 3.8). In addition to low seed input, low

numbers of perennial grass seed in the seed banks may reflect seed loss and reductions in

seed numbers from seed rain to the seed bank have been reported elsewhere (Schott and

Hamburg 1997; Rabinowitz and Rapp 1980). Such losses may be due to various factors such

as senescence and loss to deep soil, germination as well as to the activity of granivores and

disease (Kemp 1989). Insect and microbial attack have been suggested as a cause of reduced

C. ciliaris seed in the seed bank (Hacker 1989).

141

Low seed densities of perennial grasses and high densities of other species groups have

important implications for perennial grass establishment. Low numbers of perennial grass

seed suggest that these species may not be present to germinate and establish when sites

become available (McIvor 1987; refer McIvor and Gardener 1994). Some perennial grass

species may be ‘extinction-prone’ due to variable seed inputs and transient seed banks

(O’Connor 1991). Clearly, the dominance of C. ciliaris in the standing vegetation of both

communities and Astrebla at Moorrinya indicate that, despite low abundance in the seed

bank, these species have successfully recruited in the past. However, their colonizing ability

in the present environment remains unknown. The relevance of low seed numbers for the

establishment of other perennial grasses in these communities is also unclear.

An additional consequence of the composition of these seed banks is that perennial grass

establishment may be suppressed by competition from annual species. In these communities

most seeds germinate on the first rainfall event of the wet season and there is intense

competition within the seedling stand (McIvor and Gardener 1991). While perennial grasses

are good competitors once established, evidence from many studies indicates that perennial

grasses have difficulty competing with annual grasses as seedlings (refer Brown and Rice

2000). The outcome of competitive interactions depends on the species and densities

involved and environmental conditions. McIvor and Gardener (1991) stated that, while few

perennial grass seedlings would be able to survive in very competitive situations since they

do not have superior growth rates or other advantages during their early growth phase, they

could out-compete other plant groups such as legumes if competition is for light and not

water. Studies of competitive interactions between Astrebla lappacea and the annual grass

Iseilema led Orr and Evenson (1993) to conclude that Astrebla could be out-competed by

Iseilema at high densities but that forbs were unlikely to achieve densities high enough to

prevent Astrebla recruitment. Cenchrus ciliaris seedlings are believed to be poor competitors

against established plants (Hacker 1989; McIvor 2003). However, I am unaware of any work

investigating the competitive ability of C. ciliaris seedlings compared with other seedlings.

The importance of seed banks in the colonization of gaps varies between communities

(Thompson 1992). In some situations, for example where adult plants are capable of

sprouting after fire, seedling recruitment from the seed bank may be unimportant for

determining vegetation composition (Auld and O’Connell 1991). In grasslands, small gaps

tend to be filled by vegetative growth of surrounding plants, although seeds seem to be

crucial for the colonization of large gaps (Thompson 1992). The effects of fire on seed banks

may be very important where perennial grasses senesce or suffer mortality and have to be

replaced by juveniles (Ernst 1991). In terms of fire-induced changes in perennial grassland

142

composition, the effects of fire on soil seed banks may be less important than fire effects on

established plant competition and establishment site availability: unless sites become

available for establishment, fire impacts on the soil seed bank will be irrelevant. In addition,

seed banks may be strongly influenced by climate, particularly rainfall variability, and its

effects on seed production inputs and losses (refer Morgan 1998; Orr 1998). Consequently,

in these communities, climate rather that fire, may be the dominant influence on soil seed

banks and fire will play a more important role in influencing vegetation composition via its

effects on establishment site availability.

143

3.8. DOES FIRE PROMOTE FLOWERING IN CENCHRUS CILIARIS?

3.8.1. INTRODUCTION

Fire may result in shifts in the composition of plant communities via effects on flowering

and seed inputs. For example, the abundance of Bromus japonicus, an invasive grass in

North American mixed-prairie communities, may be altered by changes in the fire regime

since seed production in this species is reduced by fire (Whisenant 1990b). Species differ in

their flowering responses to fire (Glenn-Lewin et al. 1990), varying from being unaffected

by burning to near obligate associations (Bond and van Wilgen 1996) (section 3.2.2.5). Fire-

stimulated flowering is very common in grasses (Bond and van Wilgen 1996), although both

increased and decreased flowering in grasses have been associated with fire in overseas

studies (Daubenmire 1968; Glenn-Lwein et al. 1990). Fire has been reported to promote

flowering in grasses (Scanlan 1980) and forbs (Lunt 1994) in Australian grasslands.

Season of burning may have a significant influence on flowering responses (Glenn-Lewin et

al. 1990). For example, flowering of grasses of wet prairies in South Florida is promoted by

burning in the growing season but not by burning in the dormant season (Main and Barry

2002). The objective of this study was to investigate the effects of season of burning on C.

ciliaris flowering in two C. ciliaris-dominated grasslands. While flowering in C. ciliaris has

been reported to be enhanced by fire (L. Baker, pers. comm., cited in Humphries et al. 1991),

little is known about how the timing of burning may affect this response. The flowering

response of a native perennial grass, Astrebla spp., to season of burning was also

investigated at one site where this species was co-dominant.

3.8.2. METHODS

3.8.2.1. Measurements

Flowering of C. ciliaris at Dalrymple and C. ciliaris and Astrebla spp. at Moorrinya was

monitored after the implementation of burning treatments at the two sites (section 3.3). The

tagged plants used to monitor season of burning effects on plant survival were used for this

study (section 3.4). The presence of inflorescences was recorded as plants were tagged in

144

April and June 1999. The number of inflorescences per tagged plant was counted after the

fires, in January and June 2000 at Dalrymple and in February 2000 at Moorrinya. These

sampling times were chosen for logistic rather than ecological reasons.


The effects of season of burning on the numbers of inflorescences per tagged plant and the

percentage of tagged plants in flower at Dalrymple were analysed using a repeated measures

ANOVA. Inflorescence counts were log transformed and percentage flowering data arcsin

transformed prior to analysis. Means were compared using the protected LSD test at the 5%

significance level and back-transformed means are presented. Data for small and large plants

were also analysed separately using the same methods. Season of burning effects on the

numbers of inflorescences per tagged plant and the percentage of tagged plants in flower at

Moorrinya were investigated using one-way ANOVA for each species separately. Data were

transformed and reported as above. Relationships between plot fire intensity and numbers of

inflorescences and between fire intensity and percentage of flowering plants were

investigated using linear regression. All analyses were performed using Genstat (2001).

3.8.3. RESULTS

No effects of season of burning on C. ciliaris flowering were found at Dalrymple for either

small or large plants (P > 0.05). Numbers of inflorescences per plant and the percentage of

plants in flower were similar between the two sampling times and only the January data are

presented for the two size classes combined (Figures 3.18. and 3.19). More large plants

flowered (53%) than small plants (14%) and large plants had more inflorescences (5.1 ± 1.05

inflorescences/plant) than small plants (0.3 ± 0.10 inflorescences/plant). There was little

difference in the percentage of plants in flower when the plants were tagged in 1999 (30%)

and when they were re-surveyed in January (32%) and August (29%) 2000.

At Moorrinya, C. ciliaris plants in the late dry season burn treatment had more

inflorescences than plants in the early burn treatment (P < 0.05) but plants in both burning

treatments had similar numbers of inflorescences as unburnt (control) plants (Figure 3.18).

There were no burning treatment effects on numbers of Astrebla inflorescences per plant (P

145

> 0.05) (Figure 3.18) or on the numbers of C. ciliaris or Astrebla plants in flower (P > 0.05)

(Figure 3.19). Flowering was similar in the two species. In February, 30% of C. ciliaris

plants were in flower and plants averaged 2.1 ± 0.39 inflorescences/plant while 44% of

Astrebla plants were in flower and plants averaged 2.4 ± 0.31 inflorescences/plant. The

percentage of plants in flower in February 2000 was lower than in April/June 1999 when the

plants were tagged. At this time 54% of C. ciliaris plants and 98% of Astrebla plants were in

flower.

No relationships between plot fire intensity and number of inflorescences/plant or percentage

of flowering plants were found for C. ciliaris or Astrebla plants (P > 0.05).

(Figure

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

Control Early Late

Mea

n nu

mbe

r of i

nflo

resc

ence

s/pl

ant

a

a

(a) Dalrymple C. ciliaris. (b) Moorrinya C. ciliaris (c) Moorrinya Astrebla spp.

b

b

Figure 3.18. Mean number of infloresc

January 2000) and (b and c) Moorrinya

February 2000) in control (no burn), ear

Different lower case letters denote signi

a

Control

Burning

a a

ences p

(C. cil

ly dry s

ficantly

a

a

Early Late Control Early

treatments

er tagged plant at (a) Dalrymple (all p

iaris and Astrebla spp. plants respect

eason fire and late dry season fire tre

different means within graphs (P < 0.

a

Late

lants in

ively in

atments.

05).

146

0

20

40

60

80

100

Control Early Late Control Early Late Control Early Late

P (%

)

Figure 3.19. Mean percentage of tagged plants in flower at (a) Dalrymple (all plants in

January 2000) and (b and c) Moorrinya (C. ciliaris and Astrebla spp. plants respectively in

February 2000) in control (no burn), early dry season fire and late dry season fire treatments.

3.8.4. DISCUSSION

Cenchrus ciliaris is considered to be favoured by fire (Humphries et al. 1991; Lazarides et

al. 1997; Butler and fairfax 2003). However, the fires in this study appeared to have little

effect on flowering in C. ciliaris. The lack of flowering response to fire in Astrebla was also

surprising given that fire-stimulated flowering in this species has been reported (Scanlan

1980). This lack of response may be due to the low intensity of the fires. Alternatively, or in

addition, sample variability may have limited the power of statistical tests. Inflorescence

number per plant was highly variable with many plants producing no flowers while a few

produced many. Although flowering in C. ciliaris at Moorrinya was greater with the higher

intensity, late dry season fires than with the early dry season fires, the treatment effects at

this site do not appear to be related to fire intensity and are difficult to explain. No

relationships between plot fire intensity and flowering were detected and the greater

flowering of C. ciliaris in the late dry season fires may simply be a consequence of the

variability in flowering and the small sample size.

The interactions between fire, flowering and rainfall are important in seasonally dry

erce

ntag

e of

tagg

ed p

lant

s in

flow

er

(a) Dalrymple C. ciliaris. (b) Moorrinya C. ciliaris (c) Moorrinya Astrebla spp.

Burning treatments

147

ecosystems where rainfall patterns are a major factor influencing plant growth (Orr et al.

1991). It is likely that rainfall plays a major role in determining flowering in the C. ciliaris

grasslands investigated here and that the effects of fire are of secondary importance. It had

been reported that fire promotes seed production in C. ciliaris, particularly when fire was

followed by rain (L. Baker cited in Humphries et al. 1991) and Bosch and Dudzinski (1984)

noted that seed head production in C. ciliaris in central Australia was determined by rainfall.

The responses of Astrebla to burning have been found to differ depending on rainfall, with

increased seed head density on burnt plants at sites experiencing above average rainfall

(Scanlan 1980). Although overall wet-season rainfall was above average (section 3.3.5), the

soil moisture status over the growing season at the experimental sites is unknown.

Consequently, any moisture limitations to flowering are also unknown. Flowering in both C.

ciliaris and Astrebla was lower in 2000 than in 1999. However, the factors influencing

flowering responses are not clear. The effects of fire on flowering in grasses in North

American tallgrass prairies have been found to vary between years and between habitats

(Glenn-Lewin et al. 1990). More detailed studies are required to determine the effects of

growing season rainfall and other site differences on flowering responses to fire in these C.

ciliaris-dominated grasslands.

The timing of fire may have a significant influence on the degree of subsequent flowering in

grasses (Glenn-Lewin et al. 1990). The effects of season of burning on flowering and seed

inputs are critical in determining potential recruitment patterns in these grasslands.

Unfortunately, in this study it was not possible to monitor season of burning effects of the

flowering of native species in detail due to lack of resources. Studies are needed to determine

season of burning effects on flowering of C. ciliaris and local native grasses in these

communities.

It is also important to determine the effects of fire on viable seed inputs. Although

inflorescence production may be a good indication of seed production (e.g. as for Astrebla,

Orr 1998), relationships between flowering and seed inputs are not always positive. For

example, Bosch and Dudzinski (1983) reported that whereas severely defoliated C. ciliaris

plants produced more seed heads than non-defoliated plants, non-defoliated plants produced

more seeds per seed head, more than compensating for their lower seed head numbers.

Consequently, burning treatment effects on flowering may not be reflected in viable seed

production. Again, post-fire rainfall is likely to be a dominant factor determining viable seed

inputs since rainfall conditions during seed set influence seed viability (Orr 1998). Further

work is required to determine how season of burning affects viable seed inputs by C. ciliaris

and native species in these grasslands.

148

3.9. EFFECTS OF SEASON OF BURNING ON SEEDLING EMERGENCE

PATTERNS IN CENCHRUS CILIARIS-DOMINATED GRASSLAND

3.9.1. INTRODUCTION

Fire may play a major role in shaping vegetation composition via its effects on seedling

recruitment patterns. Flushes of seedling emergence after fire are commonly reported for

both herbaceous (Shaw 1957; Tothill 1969; refer Tyler 1995) and woody (Purdie and Slayter

1976; Williams 2000) species, although decreases in seedling emergence after fire have also

been found (Mott and Andrew 1985b; Glenn-Lewin 1990; Tyler 1994) (section 3.2.2.5).

The importance of seedling dynamics for vegetation composition varies, from being

fundamental in annual plant communities, to being relatively unimportant much of the time

in many perennial plant communities where the persistence of resident plants means that

recruitment events are rare (Lauenroth and Aguilera 1998). However, the composition and

establishment success of seedling cohorts may become very significant in determining

vegetation composition in perennial systems when disturbances free up resources creating

opportunities for recruitment of new individuals.

Seedling recruitment requires a source of viable propagules as well as microsites with

specific features that permit seed germination and seedling growth (Harper et al. 1965). Fire

may result in significant shifts in vegetation composition by altering both these factors. Fire

alters germinable seed availability directly, by killing seed and affecting dormancy

mechanisms, and indirectly, via effects on flowering (section 3.2.2.5). Fire also alters

establishment site availability. The environmental changes produced by fire may stimulate

seedling recruitment as competition from established plants is reduced, resources such as

light, water and nutrients are increased, allelopathic influences are decreased and seed and

seedling predation is reduced (Tyler 1995; Bond and van Wilgen 1996) (section 3.2.2.4).

Alternatively, burning may result in fewer establishment sites. For example, moisture

conditions may be less favourable for seedlings in burnt compared with unburnt sites

(Christensen and Muller 1975).

The importance of fire in promoting seedling emergence varies between communities. In

some communities, for example chaparral shrublands in North America (Tyler 1995),

significant seedling establishment occurs only following fire (Noble 1989). In other

communities, fire is less important since seedling emergence and survival are primarily

149

controlled by climate, particularly rainfall (Lauenroth et al. 1994). The importance of

seedling recruitment for post-fire recovery also varies, depending on both the fire regime and

vegetation characteristics. For example, seedling recruitment is often of little importance in

temperate Australian perennial grasslands after single fires. Single fires in these grasslands

do not promote mass seed germination of most perennial species, probably because of small

soil seed banks (Lunt and Morgan 2002) and also because the perennial grasses generally

recover vegetatively from fire (Morgan 1999). In contrast, where fire regimes result in major

plant mortality, seedling recruitment may be the only means of re-colonization.

The season of burning may influence seedling recruitment patterns since both the intensity

and timing of fire influence the availability of germinable seed and establishment sites.

Variations in fire intensity and soil heating can affect recruitment since seeds of different

species have different tolerances to heating and different germination requirements (Tyler

1995). For example, high intensity fires may result in flushes of legume germination since

these fires are more effective than low intensity fires in breaking down hard-seededness

(Mott 1982). High intensity fires are important in legume germination in dry sclerophyll

jarrah forests in Western Australia where native legume species rarely germinate after low

intensity fires (Shea et al. 1979). In addition, fire intensity influences micro-site conditions.

The degree of litter removal and mortality of established plants will influence establishment

site availability and seedling recruitment success. The timing of fire in relation to vegetation

condition also influences the effects of fire on seedling recruitment patterns. Burning annual

Sorghum grassland at the commencement of the wet season after the majority of seeds have

germinated but prior to flowering and seed set may drastically reduce future Sorghum

abundance by eliminating the source of seed (Stocker and Sturtz 1966). Conversely,

Sorghum abundance may be unaffected by dry season fires (Andrew and Mott 1983). In

perennial grasslands, early dry season fires may result in reduced seedling emergence by

destroying seed still held in seed heads or on the soil surface (Walker et al. 1983) while late

dry season fires may have little impact since seed will be buried and protected from fire by

this time.

The objective of the two studies reported here was to investigate the effects of season of

burning on seedling emergence patterns in C. ciliaris-dominated grassland. In the first study,

the numbers of grass and forb seedlings emerging early in the growing season after the

implementation of burning treatments was assessed. In the second study, the emergence of

grass seedlings was surveyed over the duration of the second growing season after the fires.

150

3.9.2. METHODS

3.9.2.1. Seedling monitoring sites

The effects of different season of burning on seedling recruitment in a C. ciliaris-dominated

grassland were investigated using plots within the main experimental area at Dalrymple

(plots one to nine, section 3.3). Eight metal quadrats were randomly positioned within the

inner 18 m by 18 m zone of each plot in December 1999. Each quadrat, measuring 50 cm by

50 cm, was made of 10 cm by 10 cm metal mesh such that each quadrat formed a five by

five grid of 10 cm by 10 cm ‘cells’. Quadrats were pegged to the ground and marked with 1

m high bamboo stakes to aid relocation.

3.9.2.2. Study 1: Seedling recruitment survey

The numbers of grass and forb (legume and other forb) seedlings in each cell were recorded

in December 1999, five weeks after the late dry season fires. Seedlings were identified to

species level where possible.

Grass and forb seedling data were analysed using one-way ANOVA. Sub-sample (quadrat)

seedling totals were square root transformed prior to analysis. Means were compared using

the protected LSD test at the 5% significance level. The reported treatment means have been

back-transformed and converted to seedling numbers/m2. Correlations between forb and

grass seedlings were investigated at the cell and quadrat level. Linear regression was used to

investigate relationships between plot fire intensity and plot seedling numbers. All analyses

were performed using Genstat (2001).

3.9.2.3. Study 2: Grass seedling recruitment and survival

To monitor grass seedling recruitment in more detail, four cohorts of grass seedlings were

tagged over the 2000-2001 growing season (in December 2000 and January, February and

March 2001) and their survival assessed at the end of the growing season in June 2001. The

metal mesh quadrats used in the previous study were used. Grass seedlings found in the nine

central cells of each quadrat were individually tagged by pushing a partly unwound paper

clip in beside the seedling. Different coloured clips were used to identify each cohort. It was

impractical to tag all seedlings when they occurred at high density. Therefore, on each

151

survey day only one seedling was tagged in a cell when it contained more than four untagged

seedlings. However, all seedlings were counted. Unfortunately, it was not possible to

determine the genus of many seedlings due to their small size and attempts at identification

were abandoned.

Cumulative seedling emergence was estimated for each survey date. In cells in which not all

seedlings were tagged, new individuals could not be distinguished from those previously

counted. In these cases, only seedlings in excess of the number that had been reported for the

previous sampling were added to the count. Seedling numbers were underestimated in this

study since emergence and death of seedlings may have occurred between sampling dates.

At the end of the growing season the quadrats were re-surveyed to assess the fate of the

seedlings. The presence or absence of all tags and seedlings was recorded. The status of

seedlings (dead or alive) could not be accurately assessed, since most herbaceous vegetation

had begun to brown off, and seedlings were counted as surviving if they were present.

Consequently, survival reported here is ‘apparent’ survival. In addition to the total number of

surviving seedlings, the numbers of surviving seedlings from each cohort were assessed.

Seedlings that could not be assigned to a cohort (untagged seedlings and seedlings with

missing tags) were excluded from these counts.

Burning treatment effects on cumulative seedling emergence counts were analysed using a

repeated measures ANOVA procedure. The cumulative seedling emergence counts were also

analysed for each sampling day separately to check homogeneity of variances. Burning

treatment effects on the total numbers of emerged, dead and surviving seedlings were

determined using one-way ANOVA. Prior to the analyses, sub-sample (quadrat) totals were

log transformed. Means were compared using the protected LSD test at the 5% significance

level and back-transformed means converted to counts/m2 are presented. Seedling survival

data were also expressed as percentages of emerged seedlings. Burning treatment effects on

the percentage seedling survival per plot were determined using one-way ANOVA with

percentages being arcsin transformed prior to analysis. Means were compared using the

protected LSD test at the 5% significance level and back-transformed means converted to

counts/m2 are presented. Burning treatment effects on the emergence and survival of

seedlings in individual cohorts could only be assessed for the December cohort because

seedling numbers in the other cohorts were too small. Data from the December cohort were

analysed using one-way ANOVA as above.

Relationships between forb and grass seedling emergence in 1999 and relationships between

152

grass seedling emergence in 1999 and 2000-2001 were investigated using linear correlation.

Relationships between estimated plot fire intensity (kW/m) and plot seedling emergence

were investigated using linear regression. Relationships between flowering of tagged C.

ciliaris plants (section 3.8) and grass seedling emergence were also investigated using linear

regression. All analyses were performed using Genstat (2001).

3.9.3. RESULTS

Season of burning had a significant effect on forb seedling emergence. Five weeks after the

late dry season fires, mean forb seedling number was higher in the early dry season burn

treatment than in the other treatments (P < 0.05) (Figure 3.20). The same trends were found

for grass seedlings, although treatment means were not significantly different (P > 0.05)

(Figure 3.20).

0

10

20

30

40

50

60

70

80

Control (no fire) Early fire Late fire

Mea

n nu

mbe

r of s

eedl

ings

/m2

Burning treatments

a

b

a

Figure 3.20. Mean number of forb seedlings (solid bars) and grass seedlings (unfilled bars)

in control, early dry season and late dry season fire plots at Dalrymple in December 1999.

Different lower case letters denote significantly different forb seedling numbers.

Relationships between fire intensity and seedling emergence are unclear since the uneven

spread of fire intensity values prevents valid assessment of any relationship (Figure 3.21).

153

While there was a significant linear relationship between plot fire intensity and forb seedling

numbers (P < 0.05), no relationship was found when the high fire intensity plot, which had

high leverage, was excluded from the analysis. No relationship between fire intensity and

grass seedling numbers was detected (P > 0.05).

0

20

40

60

80

100

120

140

160

0 500 1000 1500 2000 2500 3000 3500

Plot fire intensity (kW/m)

Mea

n nu

mbe

r of s

eedl

ing/

m2

Figure 3.21. Relationships between estimated plot fire intensity and mean number of grass

(∆) and forb (■) seedlings/m2. Significant linear regression shown for forb seedling numbers

(Forb seedling number = 11.2 +0.027*Fire intensity, adjusted R2 = 0.50, P < 0.05).

Overall, seedling emergence was patchy with 19% of quadrats containing no seedlings. On a

quadrat basis grass seedling densities ranged from 0 to 216 seedlings/m2 (mean of 49 ± 7.0

seedlings/m2) while forb seedling densities ranged from 0 to 212 seedlings/m2 (mean of 36 ±

6.0 seedlings/m2). There was little correlation between the occurrence of forb and grass

seedlings in 1999 (r = 0.20 and 0.42 at the cell and quadrat levels respectively).

Almost half of the forb seedlings appeared to be species of Indigofera. These legumes

appeared to be favoured by early dry season fires with means of 1, 31 and 9 seedling/m2 in

unburnt, early dry season and late dry season fire treatments respectively. However, these

treatment means were not significantly different (P > 0.05). Seedlings of Boerhavia,

Portulaca and Sida species were also identified. Most grass seedlings could not be identified

to genus level.

154

No burning treatment effects on total grass seedling emergence over the 2000–2001 growing

season were detected (P > 0.05) (Figure 3.22). However, numbers of surviving seedlings

were higher for seedlings that emerged in the late dry season fire treatment than in the

control or early dry season fire treatments (P < 0.05) (Figure 3.22). No treatment effects on

the proportion of surviving seedlings were found (P > 0.05).

0

10

20

30

40

50

60

70

80

Control (no fire) Early fire Late fire

Mea

n nu

mbe

r of s

eedl

ings

/m2

Burning treatments

b

a

a

Figure 3.22. Mean number of grass seedlings (seedlings/m2) found in control (no burn),

early dry season and late dry season fire treatments showing seedling status (dead ■ and

surviving □) at the end of the 2000-2001 growing season (June 2001). Different lower case

letters denote significantly different (P < 0.05) means for surviving seedlings.

There was a significant time effect on emergence with most seedlings (85%) emerging by

the December survey date. By January and February a further 8% and 4% of seedlings had

emerged respectively. The small numbers of seedlings emerging after December meant that

trends in cumulative emergence with fire treatments remained constant over time.

Season of burning effects on cohort seedling survival were investigated only for the

December cohort because of low seedling numbers in other cohorts. Treatment effects on

December cohort seedlings reflected the overall results with greater numbers surviving in the

late dry season fire treatment (P < 0.05). The percentage of seedlings surviving to June 2001

increased with cohort age: surviving seedlings made up 45%, 30%, 21% and 20% of

seedlings in the December, January, February and March cohorts respectively.

155

There was no relationship between numbers of grass seedlings emerging in the 2000-2001

growing season per plot and mean number of inflorescences of C. ciliaris tagged plants

earlier in the year (P > 0.05). No relationship between plot fire intensity and 2000-2001 grass

seedling numbers was detected (P > 0.05).

As in the previous growing season, grass seedling emergence was patchy with 26% of

quadrats containing no seedlings. On a quadrat basis seedling density ranged from 0 to 1711

seedlings/m2 (mean of 89 ± 26.7 seedlings/m2). There was no relationship between grass

seedling emergence in 1999 and 2001 (r = 0.06 and 0.31 at the cell and quadrat levels

respectively).

The tagging method was moderately successful with 79% of tags still in place at the end of

the study. About 60% of the lost tags were from cells in which no seedlings survived so their

loss did not affect seedling assessment.

3.9.4. DISCUSSION

Fire may alter seedling recruitment patterns and faster and greater seedling emergence after

fire is often reported (Tothill 1969; Tyler 1995; Williams 2000). In this study, seedling

emergence tended to be higher in burnt than in unburnt plots. However, treatment differences

were statistically significant for forbs only. Given the patchy distribution of seedlings, the

sampling regime used may have been insufficient to adequately describe seedling emergence

patterns. A much greater sampling effort may have detected more treatment effects.

However, it is also likely that the lack of major burning treatment effects on seedling

emergence reflects the relatively low intensities of the fires.

It is not surprising that forb emergence was promoted by burning since about half the forb

seedlings were legumes. Fire is known to promote legume germination via heat effects on

the seed coat that break seed dormancy (refer Auld and O’Connell 1991) and flushes of

legume germination after burning have been reported (’t Mannetje et al. 1983; Robertson et

al. 1999) (section 3.2.2.5). The effect of season of burning was probably related to

differences in fire intensity, with more forb seedlings emerging in the more intense, early dry

season burn treatment. While high intensity fires can kill seed (Auld and O’Connoll 1991),

they may be more effective than low intensity fires in breaking down hard-seededness (Mott

156

1982). However, no convincing relationship between plot fire intensity and forb seedling

emergence was found. A more even spread of plot fire intensity values is needed to

investigate fire intensity-emergence relationships. Although the increase in forb emergence

with early dry season burning may have been due, in part, to the increased germinability of

legume seeds, fire may also alter emergence patterns via effects on micro-site conditions. It

is not known how much of the forb responses to burning were due to direct effects on seed

germinability versus the creation of favourable micro-sites.

Burning had no detectable effects on grass seedling emergence. Trends in grass seedling

emergence with season of burning were not statistically significant or consistent between

years. The greater numbers of surviving grass seedlings in the late dry season burn treatment

in June 2001 was simply due to greater emergence in this treatment rather than to fire effects

on seedling survival since there was no treatment effects on the proportion of grass seedlings

surviving.

Seedling emergence patterns following fire may reflect burning effects on seed and

establishment site availability. The trend of greater grass seedling emergence with burning

than without burning in the first growing season after the fires perhaps indicates that positive

fire effects on establishment sites out-weighed any negative effects on sites or seed

availability. Species variations in seed burial mechanisms may result in significant shifts in

the relative abundance of species among seedling cohorts emerging after fire (Peart 1984).

However, since grass seedlings could not be identified, it was not possible to investigate

burning effects on seed of individual species. Grass seedling emergence patterns in the

second growing season may reflect burning effects on flowering. Since perennial grasses in

these communities do not form persistent seed banks (see Silcock and Smith 1990 for C.

ciliaris seed banks), most seedlings in the second season would have emerged from seed

produced after the fires. Relationships between flowering and seedling emergence could not

be investigated in detail since the flowering data was for C. ciliaris plants only (section 3.8)

whereas grass seedling counts included other species. Although no relationship between plot

means of inflorescences per C. ciliaris plant and seedling numbers was found, trends in

treatment means support the hypothesis that inflorescences per C. ciliaris plant (section

3.8.3) and grass seedling numbers tended to be higher in burnt than in unburnt plots.

However, differences in treatment means were not statistically significant and it is not

possible to identify factors influencing grass seedling emergence patterns.

There was no correlation between forb and grass emergence sites. This may reflect the

patchy distribution of seed (Bigwood and Inouye 1988; Kemp 1989). It may also indicate

157

that sites favourable for forb emergence were not necessarily favourable for grass

emergence. The effects of particular microsites on germination and emergence vary

depending on the species involved since seeds often possess highly specific germination

requirements (Harper et al. 1965). Differences in seed size and shape can affect seed-soil

contact (Evans and Young 1970; Winkel et al. 1991) and therefore the same micro-site may

present different moisture conditions for seeds of different species. Differences in seedling

morphology can also influence the suitability of a particular micro-site for emergence (Gross

and Werner 1982). There was also no relationship between grass emergence sites in the first

and second growing seasons after the fires. This may reflect differences in patterns of seed

availability and/or micro-site conditions between the two seasons. In addition, there may be

inter-year variations in the degree to which particular sites are favourable for seedling

establishment (Fowler 1988).

Seedling survival is influenced, in part, by both the spatial and temporal patterns of seedling

emergence. Where seedlings emerge in high densities there may be high mortality and

McIvor and Gardener (1991) concluded that few perennial grass seedlings would survive

such competitive situations. However, the reverse situation has also been observed, where

seedling survival is greater when there are other seedlings nearby (Fowler 1988), the likely

explanation being that the presence of neighbours reflected the favourableness of the site

(Fowler 1988).

Temporal patterns of seedling emergence may also influence seedling survival. Earlier

germinating seedlings often have a higher probability of survival (Hamrick and Lee 1987;

McIvor 1987; Fowler 1988). However, this is not always the case. Rapid germinators may

have a competitive advantage if conditions remain favourable and these seedlings have time

to grow deeper roots, and hence, are more drought tolerant than later germinating individuals

(refer Fowler 1988). However, slow germinators, which may fail to germinate during short

favourable periods, could benefit in the long run if they are able to germinate after a

catastrophe has killed the rapid germinators (Grubb 1977).

Interactions between competing seedlings are of particular interest where a change in

dominance is desired. The results from this study suggest that forb competition is unlikely to

affect grass seedling emergence and earlier emerging seedlings have a better chance of

survival than later emerging seedlings. However, the outcome of inter-species competition

between grass seedlings is unknown. Although C. ciliaris seedlings are believed to be poor

competitors against established plants (Hacker 1989; McIvor 2003), little is known about

competitive interactions between C. ciliaris seedlings and seedlings of other species. This is

158

a key issue that requires investigation.

The results from these studies provide little information regarding how fire may affect

community composition since it was not possible to identify seedlings and follow the fate of

individual species. An understanding of fire-induced emergence patterns is useful for

predicting potential vegetation change. However, it is important to recognise that micro-sites

favourable for germination and emergence may not necessarily be conducive to seedling

establishment (Schupp 1995). For example, surface cover may be important in determining

germination and emergence success (section 3.6), whereas other factors, such as distance

from nearest neighbours may be more important in determining growth and reproductive

output (Gross and Werner 1982). Although the fate of the surviving seedlings is unclear, it is

likely that few will persist and become established. In perennial grasslands, the survival and

establishment of seedlings is greatly influenced by competition from established plants for

water and nutrients (Cook 1980) and removal of this competition is a key factor in promoting

seedling survival (Cheplick 1998). The low intensity fires in this study had little effect on the

competitiveness of established plants since few plants were killed (section 3.4), although

there was a decline in perennial grass cover with burning (section 3.10). McIvor and

Gardener (1981) concluded that considerable disturbance was required for C. ciliaris

establishment and that it was likely that the temporary reduction in above ground

competition by burning had no beneficial effect. However, C. ciliaris seedling establishment

after fire has been observed (Back 2001) and Cook (1984) reported that, although emergence

of C. ciliaris was lower on burnt than on unburnt plots, seedling survival was higher. In the

context of using fire to reduce the abundance of C. ciliaris, an understanding of season of

burning effects on C. ciliaris and native species recruitment is essential.

Seedling establishment patterns are difficult to predict since they are a function of interacting

abiotic factors, such as rainfall distribution and soil type, and biotic factors, such as

competition and herbivory (Cheplick 1998). Seedling recruitment is sensitive to soil water

availability and seedling recruitment patterns are influenced by soil texture and rainfall

(Lauenroth et al. 1994). Climate will have an over-riding effect on seedling establishment in

these C. ciliaris-dominated communities with the timing of the opening rains and the

adequacy of subsequent rainfall being a major influence on seedling recruitment (McIvor

and Gardener 1991). Further work is required to investigate the influence of fire on the

establishment of C. ciliaris relative to native species in relation to rainfall regimes.

159

3.10. EFFECTS OF SEASON OF BURNING ON HERBACEOUS COMMUNITY

COMPOSITION OF CENCHRUS CILIARIS-DOMINATED GRASSLAND

3.10.1. INTRODUCTION

Fire is an important tool for eliminating or reducing the abundance of unwanted species. It is

one of the few tools available that can be used economically and effectively over large areas

and the strategic application of fire regimes has been employed to manage invasive woody

and herbaceous species (section 3.2.3).

The efficacy of fire in controlling invasive plants depends on the fire regime and on the

physiological and morphological properties of both the native community and the invading

organism (Christensen and Burrows 1986). Fires can be imposed at different frequencies,

intensities and in different seasons and the responses of vegetation to fire vary with fire

regime (section 3.2.3). In addition, the ecological consequences of particular fire regimes

depend on the species present and stage of life cycle when exposed to fire (Gill et al. 1990).

Fire may result in compositional changes where the resident plants differ in size, vigour,

morphology and/or life strategy. Post-fire conditions such as the distribution of rainfall and

grazing also influence the responses of vegetation to burning. For example, fire effects on

species richness may be dependent on the prevailing moisture conditions (Walker and Peet

1983).

The season of burning can have a marked effect on the structure and composition of

grassland vegetation (Tothill 1971; Hodgkinson et al. 1984; Collins and Gibson 1990; Howe

1994) since the effects of fire vary with seasonal changes in fire characteristics and

vegetation condition. Fire intensities vary seasonally. Generally, late dry season fires are

more intense than early dry season fires since fuel moistures are normally lower at the end of

the dry season and weather conditions at this time of year promote more intense fires (Gill et

al. 1996). Fire intensity influences the amount of biomass consumed (Williams and Cook

2001), plant mortality (Williams J. R. 1995), and germination responses (Shea et al. 1979)

and seasonal changes in intensity may result in fires in different seasons having very

different effects (section 3.2.3.1). The effects of fire also depend on the condition of the

vegetation at the time of burning (section 3.2.3.2). In tropical savannas, perennial grass

plants become dormant at the end of the growing season, translocating resources to their

roots. Consequently, individuals are likely to be more susceptible to burning at the start of

the dry season, before they have fully senesced, than at the end of the dry season when loss

160

of above ground dry matter is of little consequence. The timing of fire in relation to

flowering and the location of seed can also be critical to the response of vegetation to fire

(section 3.2.3.2). In addition, other seasonal factors may influence vegetation responses to

burning. For example, differences in the timing of soil cover removal by fire may have

significant consequences: by removing cover immediately prior to wet season rains, late dry

season fires may result in greater soil erosion and nutrient loss than early dry season fires

(section 3.2.3.1).

The aim of this study was to investigate the effects of season of burning on the herbaceous

species composition of C. ciliaris-dominated grassland. Of particular interest was whether

fire could be used to reduce the abundance of C. ciliaris. While C. ciliaris is a fire-adapted

species, changes in fire regime may alter plant competitive interactions and the strategic use

of fire has been suggested as a method for maintaining or restoring grasslands invaded by C.

ciliaris (Daehler and Carino 1998). The effect of C. ciliaris on fire regimes is frequently

highlighted in literature discussing it as an invasive species (e.g. Humphries et al. 1991; Low

1997). However less is known about fire regime effects on C. ciliaris dynamics. In a pastoral

context, fire has not been seen as particularly useful for maintaining C. ciliaris pastures

(McIvor and Gardener 1981; ’t Mannetje et al. 1983; Pressland and Graham 1989). In

contrast, in literature discussing C. ciliaris as an invasive species, fire is believed to promote

it (Lazarides et al. 1997; Butler and Fairfax 2003) (section 3.2.4). In this study, the effects of

single fires and season of burning on C. ciliaris cover as well as herbaceous species

composition were investigated in two C. ciliaris-dominated grasslands.

3.10.2. METHODS

3.10.2.1. Plant surveys

Plant species composition and abundance in experimental plots were assessed in April 1999

at Dalrymple and in April and June 1999 at Moorrinya, prior to implementing the burning

treatments, and again at the end of the post-fire growing season, in June and July 2000 at

Dalrymple and in August 2000 at Moorrinya (section 3.3). Each plot was searched for 10-15

minutes to identify the species present. In addition, 20 1 m2 quadrats per plot were assessed

in detail. The quadrats were located by throwing a 1 m by 1 m wire frame from five

approximately equally spaced positions along each side of the plot. For each quadrat, the

161

species present and their visually estimated percentage cover were recorded. Percentage

cover was the percentage of the quadrat covered by a species. Consequently, where species

overlapped, total cover could exceed 100%. Plants were identified to species level where

possible and specimens were collected to assist with plant identification.

Total herbaceous species richness as well as species richness within plant groups (perennial

grasses, annual grasses, sedges, legumes and other forbs (referred to as forbs) were

calculated for each plot at the quadrat (1 m2) and the plot (400 m2) scale. Not all plants could

be identified to species level. Therefore species richness includes some taxa identified only

to genus level. Consequently, actual species richness is underestimated in the reported

results. The abundance of individual species was assessed in terms of percentage cover and

relative cover (percentage cover/total cover) as well as frequency of occurrence (proportion

of 20 quadrats containing the species). Burning treatment effects on percentage cover were

assessed for those species that contributed at least 1% to total cover, averaged over all plots.

Burning treatment effects on the frequency of individual species was assessed for those

species occurring in at least eight plots. The percentage cover and relative cover of plant

groups were also calculated.


Burning treatment effects on herbaceous species richness and plant cover and frequency

measures were investigated using analysis of covariance (ANCOVA) of the 2000 data using

the 1999 data as a covariate where the covariate was effective. The covariate was considered

effective if significant at the P < 0.10 level. Where the covariate was not significant, the

2000 data were analysed using one-way ANOVA. (Burning treatment effects were

considered significant at the P < 0.05 level). Species frequency data were arcsin transformed

prior to analysis and the reported results are back-transformed means. Relationships between

fire intensity and plant cover and fire intensity and species richness were investigated using

linear regression. Data from the two sites were analysed separately and all analyses were

performed using Genstat (2001).

162

3.10.3. RESULTS

Burning increased species richness at the 1 m2 scale at both sites. At Dalrymple, total species

richness differed between treatments (P < 0.05) with greatest richness in the early burn

treatment (5.7 ± 0.18 species/m2), followed by the late burn treatment (4.3 ± 0.18 species/m2)

and the unburnt control (3.1 ± 0.18 species/m2). Similar trends in species richness with

burning treatments were found for the major plant groups (Figure 3.23). Perennial grass and

legume species richness were higher in the early burn treatment compared with the unburnt

treatment (P < 0.05) while a comparable trend was not statistically significant (P > 0.05) for

forb species richness. At Moorrinya, mean total species richness in early and late dry season

burn treatments (5.5 ± 0.25 and 5.3 ± 0.25 species/m2 respectively) were greater than in the

unburnt control (4.4 ± 0.25 species/m2) (P < 0.05). Forb and annual grass species richness

was also higher with burning at Moorrinya (P < 0.05) (Figure 3.23).

Species richness at the 400 m2 scale appeared unaffected by burning treatments at both sites

(P > 0.05). At Dalrymple, species richness ranged from 18-33 (mean of 25 ± 1.0) species/400

m2. Overall 72 taxa were found at this site including three annual grasses, 17 perennial

grasses, one sedge, 36 forbs and 15 legumes. At Moorrinya species richness ranged from 19-

43 (mean of 36 ± 1.4) species/400 m2. Overall 75 taxa were found including six annual

grasses, 16 perennial grasses, two unidentified grasses, four sedges, thirty nine forbs and

eight legumes. Lists of identified species for both sites are given in appendices 4A and 4B.

No relationships between plot fire intensity and species richness at the 1 m2 or 400 m2 scales

were detected (P > 0.05).

Burning reduced perennial grass cover and increased legume cover (P < 0.05) (Figures 3.24

and 3.25). At Dalrymple, the relative cover of perennial grasses was 97±1.6% with no

burning and decreased to 88 ± 1.6% and 93±1.6 % in the early and late dry season burn

treatments respectively (P < 0.05). At Moorrinya, the relative cover of perennial grasses was

lower in the late dry season burn treatment (71 ± 2.4%) than in the unburnt treatment (85 ±

2.4%) (P < 0.05). Relative perennial grass cover in the early dry season burn treatment (78 ±

2.4%) was not significantly different from the other treatments (P > 0.05). At Moorrinya,

legume cover was higher in the late dry season burn treatment compared with other

treatments (P < 0.05) (Figure 3.25). Relative legume cover was also higher in the late burn

treatment (9 ± 0.8 % relative cover) compared with the early burn and unburnt treatments (3

± 0.8% and 5 ± 0.8% relative cover respectively) (P < 0.05). At Dalrymple burning treatment

effects on legume cover were not significant (P = 0.09) (Figure 3.25). However, treatment

163

0

0.5

1

1.5

2

2.5

3

3.5 Moorrinya

0

0.5

1

1.5

2

2.5

3

3.5

0

0.5

1

1.5

2

2.5

3

3.5

0

0.5

1

1.5

2

2.5

3

3.5

Num

ber o

f spe

cies

/m2

0

0.5

1

1.5

2

2.5

3

3.5

Control

Dalrymple)

Early Late

0

0.5

1

1.5

2

2.5

3

3.5

Control Early Late

b

ab

a

0

0.5

1

1.5

2

2.5

3

3.5

Control Early Late

0

0.5

1

1.5

2

2.5

3

3.5

Control Early Late

Burning t

b

aa

(g)

Figure 3.23. Mean (± SE) species richness (number

dry season burn treatments for (a,b) annual gras

leguminous forbs and (g,h) legumes at Dalrymple (1

Different lower case letters denote significantly diffe

(b)

Control Early Late

baba

(d)

Control Early Late

(f)

(c)

(a

(e)

Control Early Late

Control Early Late

b

ab

a

(h)

reatments

of species/m2) in control, early and late

ses, (c,d) perennial grasses, (e,f) non-st column) and Moorrinya (2nd column).

rent means within graphs (P < 0.05).

164

0

10

20

30

40

50

60

70

80

0

10

20

30

40

50

60

70

80

Control Early Late

Burning treatm

b

aa

b

aa

))

Mea

n pe

rcen

tage

cov

er (%

)

Figure 3.24. Mean C. ciliaris percentage cover (sha

percentage cover (bar) cover at (a) Dalrymple and (b)

denote significantly different means within graphs (P

effects on relative legume cover were significant: re

early dry season burn treatment (10 ± 1.6%) than in

treatments (0.3 ± 1.6% and 4 ± 1.6% relative cover

forb and sedge cover is also presented in Figure 3.25.

Perennial grass cover declined with increasing fire

explained 49% of the variation in cover (P < 0.05) (F

significant linear relationship between perennial grass

R2 = 0.34) when the high intensity plot was omitte

relationships between fire intensity and legume cover

Perennial grasses dominated plots in terms of cover

growing season, perennial grasses contributed 93 ±

abundance (5 ± 1.5% relative cover) while annual g

less than 2% of total cover at Dalrymple. This compo

At Moorrinya, overall plot composition differed betw

contribution to cover decreased from 92 ± 1.7% t

contribution increased from 3 ± 1.4% to 10 ± 1.4%.

groups were less marked and in 2000 forbs contribu

sedges 1 ± 0.4% to total cover.

(b
(a
Control Early Late

b

aa

ents

ded portion) and total perennial grass

Moorrinya. Different lower case letters

< 0.05).

lative legume cover was greater in the

the unburnt and late dry season burn

respectively) (P < 0.05). Annual grass,

intensity. At Moorrinya, fire intensity

igure 3.26). At Dalrymple, there was a

cover and fire intensity (P < 0.05, Adj

d from the analysis (Figure 3.26). No

were found (P < 0.05).

at both sites. At the end of the 2000

1.5% cover and legumes were next in

rasses, sedges and forbs each made up

sition was similar to that found in 1999.

een the two years. The perennial grass

o 78 ± 2.4% while the annual grass

Changes in the relative cover of other

ted 6 ± 1.0%, legumes 5 ± 0.7% and

165

0

1

2

3

4

5

6

7

8

9

10

Control Early Late10

Moorrinya

0

1

2

3

4

5

6

7

8

9

10

Control Early Late

0

1

2

3

4

5

6

7

8

9

10

0

1

2

3

4

5

6

7

8

9

10

0

1

2

3

4

5

6

7

8

9

10

Dalrymple

Mea

n pe

rcen

tage

cov

er (%

)

Figure 3

(d) sedg

letters de

(b)

0

1

2

3

4

5

6

7

8

9

Control Early Late

0

1

2

3

4

5

6

7

8

9

10

Control Early Late10

b

a

Control Early Late

Control Early Late

a

(d)

(c)

e

(a)

0

1

2

3

4

5

6

7

8

9

Control Early LateControl Early Late

Burning treatments

.25. Mean (± SE) percentage cover of (a) annual grasses, (b) forbs, (c) legumes and

s at Dalrymple (1st column) and Moorrinya (2nd column). Different lower case

note significantly different means within graphs (P < 0.05).

166

On an individual species basis, only C. ciliaris cover was affected by burning. At Dalrymple,

its cover was significantly lower in the burnt treatments compared with the unburnt control

(P < 0.05) (Figure 3.24). Relative C. ciliaris cover was significantly lower in the early burn

treatment (44 ± 2.8%) compared with the late burn and unburnt treatments (57 ± 2.8% and

66 ± 2.8% relative cover respectively) (P < 0.05). At Moorrinya, C. ciliaris cover tended to

be lower in the early burn treatment compared with the late burn and unburnt treatments (P =

0.05) (Figure 3.24). Treatment effects on relative C. ciliaris cover were not detected at this

site (P > 0.05). Cenchrus ciliaris cover declined with increasing fire intensity at Dalrymple

(P < 0.05, Adj R2 = 0.41) while no relationship between C. ciliaris cover and fire intensity

was found at Moorrinya (P > 0.05). No burning treatment effects on the cover of other

species investigated were detected (P > 0.05). These species included the exotic perennial

grasses Bothriochloa pertusa, Melinis repens and Urochloa mosambicensis, native grasses

Bothriochloa ewartiana, Chrysopogon fallax, Heteropogon contortus and Enneapogon spp.,

the legume Indigofera colutea and Fimbristylis spp. at Dalrymple and the perennial grasses

Astrebla squarrosa, Dichanthium sericeum, Dichanthium setosum, Enneapogon polyphyllus,

and the annual grasses Iseilema spp. at Moorrinya.

0

10

20

30

40

50

60

70

80

90

100

0 500 1000 1500 2000 2500 3000 3500

Fire intensity (kW/m)

Per

enni

al g

rass

cov

er (%

)

Figure 3.26. Relationship between fire intensity and perennial grass cover at (∆) Dalrymple

and (■) Moorrinya. The regression line (perennial grass cover = 61.51-0.00815*fire

intensity) is shown for Moorrinya.

167

The frequencies of occurrence of most species investigated (13 species at Dalrymple and 14

species at Moorrinya) were unaffected by burning. However, at Dalrymple Indigofera

colutea was more abundant in the early dry season burn treatment (61% of quadrats)

compared with the late dry season burn and unburnt treatments (20 and 7 % of quadrats

respectively) (P < 0.05). At Moorrinya, Iseilema spp. were also more abundant in the early

dry season burn treatment, being found in 62% of quadrats compared with 41% and 40% of

quadrats in the late dry season burn and unburnt treatments respectively (P < 0.05). A forb,

Abutilon malvifolium, also occurred more frequently with burning at Moorrinya, being found

in 22% and 12% of quadrats in early and late dry season burn treatments respectively

compared with 2% of quadrats in the unburnt treatment (P < 0.05).

The species present at the two sites varied between growing seasons. At Dalrymple, 14

species were found in 1999 but not in 2000 and 12 species were found in 2000 but not in

1999. At Moorrinya, six species were found in 1999 but not in 2000 and 22 species were

found in 2000 but not in 1999. Although more species were found at Moorrinya in the post-

fire growing season, just under half of those found only in 2000 occurred in unburnt plots,

and, at both sites, many species were found in less than three plots, preventing the

investigation of any season of burning effects on their abundance.

3.10.4. DISCUSSION

Burning had minor effects on the composition of these C. ciliaris-dominated communities.

Although there was a reduction in perennial grass cover and an increase in species richness

at the 1 m2 scale at both sites, there were no major changes in composition with burning.

Other studies have also reported that fire results in little change to vegetation composition

(Gill et al. 1990; Lunt 1990; Tolhurst 1996). Generally, fire regimes in Australian tropical

savannas affect species abundance rather than cause significant changes in the species

present (Bowman et al. 1988; Lonsdale and Braithwaite 1991; Williams et al. 2003b),

although significant changes in structure with burning have been reported for tropical

woodlands (Bowman et al. 1988; Gill et al. 1990). The abundance of four species was

affected by burning. However, most species occurred too infrequently for fire effects on their

abundance to be determined. Fire plays a critical role in structuring many plant communities

(refer Tyler 1995), particularly where critical ecosystem processes such as seed germination

are controlled by fire, as in heathlands and wet sclerophyll forests (Lunt and Morgan 2002).

168

In contrast, fire may be relatively unimportant in effecting vegetation change in communities

where edaphic (Bowman 1988) or climatic factors (Williams et al. 2003b) have a large

influence on vegetation dynamics. In these savanna communities, the primary determinants

of composition and structure are variations in moisture and nutrients while fire and other

disturbances are secondary determinants (refer Williams et al. 2002).

The season of burning has been found to affect grassland composition. For example, wet

season fires resulted in significantly different forb species composition compared with dry

season fires in sorghum-dominated savannas in northern Australia (Lane and Williams

1997). Early dry season fires promoted Themeda triandra while late dry season fires

promoted H. contortus in Themeda-Heteropogon grasslands in central Queensland (Walker

et al. 1983). Overall, season of burning had only minor effects on these C. ciliaris-dominated

grasslands. In particular, legumes appeared to be responsive to season of burning. At

Dalrymple, I. colutea cover and legume species richness were higher with early dry season

burning while at Moorrinya, legume cover was higher with late dry season fires. These

results probably reflect fire intensity effects (the mean intensity of late dry season fires was

higher than that for early dry season fires at Moorrinya but the reverse occurred at

Dalrymple), although direct relationships between fire intensity and legume cover or species

richness were not detected. Legume germination is promoted by heat (Bond and van Wilgen

1996). Increases in legume emergence after fire have been reported (Williams et al. 2003b)

and germination responses have been found to be positively associated with fire intensity

(Shea et al. 1979). Legume abundance was found to be greater with early rather than late dry

season fires in a coastal eucalypt savanna in north Queensland (Williams et al. 2003b),

although no explanation for this was given. The annual grass Iseilema spp. was also affected

by season of burning, increasing in frequency of occurrence in early dry season burnt plots at

Moorrinya. Burning has been found to reduce Iseilema seedling emergence but increase

seedling tillering (Scanlan and O’Rourke 1982). It is not clear whether the higher frequency

of this species in early dry season burnt plots reflects greater seedling numbers or larger

plant size or why the effect occurred with early but not late dry season burning. This result

conflicts with the trend of lower Iseilema seedling emergence from burnt treatments in the

seed bank study (section 3.7.3). However, the seed bank result is likely to be an artefact of

seed bank determination procedures (section 3.7.4).

In addition to affecting community composition, fire may affect productivity. The reported

effects of fire on biomass production vary and both increases and decreases in post-fire

production have been reported (section 3.2.2.5). Post-fire production may be reduced, in part

because fire kills weaker plants (Tothill 1971). However, in this study the reduced cover

169

with burning is likely to be a reflection of post-fire growing conditions rather than a

consequence of plant death as few plants were killed (section 3.4). It is not clear what limited

growth at the two sites in my study. Post-fire biomass production is strongly dependent on

seasonal rainfall (Orr et al. 1991). Unfortunately, detailed weather data are not available for

the sites, although rainfall summaries for nearby locations indicate above average rainfall

over the 1999-2000 growing season (section 3.3.4.3). At Dalrymple, C. ciliaris in the

experimental area appeared ‘run down’ (see Pressland and Graham 1989) in that plants were

smaller and not as green as plants in other areas nearby, possibly reflecting nutrient

limitations to growth. At Moorrinya, moisture may have been limiting during parts of the

growing season, despite the likelihood of above average wet season rainfall (section 3.3.4.3).

Rainfall for a nearby site was above average but highly variable over time. An uneven

distribution of rainfall over time can lessen rainfall effectiveness. An alternative or additional

explanation for the reduced cover in burnt plots is that fire reduced the vigour of the

surviving plants.

Post-fire productivity may be affected by the season of burning. Burning at the end of the dry

season is considered the least injurious time since plants are dormant (West 1965) and in

northern Australia the smallest reductions in biomass after burning have been found for end

of dry season fires (Smith 1960). However, with the exception of legume cover, no

differences in plant cover between early and late dry season burning were found in this

study. Rather, reductions in perennial grass cover appeared to be related to fire intensity.

Generally the higher the intensity, the lower the following season’s perennial grass cover,

although a more even distribution of intensities is needed to clearly determine trends in plant

cover in relation to fire intensity. High intensity fires are likely to kill more perennial plants

and may reduce the growth of those remaining (Smith 1960). However, fire intensity-cover

relationships may reflect fire effects on the cover produced before the fire rather than fire

effects on plant growth. Since cover measured in the post-fire growing season includes cover

produced in the previous season, the greater carry-over of cover in less intensely burnt plots

may have contributed to their greater post-fire cover.

The issue of whether fire can be used to reduce C. ciliaris abundance was not resolved by

this study. Although burning reduced C. ciliaris cover at Dalrymple, it had no such effect at

Moorrinya. Importantly however, fire did not promote C. ciliaris cover. There is

considerable conjecture in the invasive plant literature regarding fire and C. ciliaris. The

results of this study do not support the positive fire loop model proposed (Butler and Fairfax

2003). This is not to say that fire never promotes C. ciliaris biomass. Rather, the results

show that C. ciliaris-fire interactions are complex and other factors such as site productivity

170

and rainfall may have significant effects. For example, patterns of production in burnt

compared with unburnt C. ciliaris grassland is greatly influenced by rainfall (Hamilton and

Scifres 1982).

The reduction in C. ciliaris cover at Dalrymple with burning appeared to be related to fire

intensity. Although there is some evidence to suggest that higher intensity fires killed more

C. ciliaris plants (section 3.4), it is not clear how much of this effect is due to removal or

reduced vigour of burnt plants versus the reduction in carry-over cover from the previous

season. It is not clear why C. ciliaris cover was unresponsive to burning at Moorrinya. Site

productivity may affect the accumulation of biomass after fire (Lunt and Morgan 2002). The

growth of plants at Dalrymple may have been compromised by nutrient limitations, making

these plants here more susceptible to fire than plants at Moorrinya. However, while C.

ciliaris grassland at Dalrymple appeared to be nutrient limited, results from the bioassay

study (section 3.5) suggest that nutrient availability was higher at Dalrymple than at

Moorrinya.

No burning treatment effects on the cover of the other perennial grasses in these

communities were found. Seedling emergence and growth of some grasses such as H.

contortus are known to be promoted by fire (Shaw 1957). Although no evidence of increased

H. contortus abundance with burning was found, an increase in H. contortus seedlings may

have been difficult to detect since the seedlings may not have been large enough to be

identified and/or may have contributed little to cover. It is possible that changes in vegetation

composition due to burning may become evident in later seasons if new individuals

established after fire.

Although burning did not result in major compositional change in these grasslands, it did

result in increased abundance of three species, in terms of frequency of occurrence, as well

as an increase in species richness at the small scale (1 m2). This contrasts with results from a

trial in central Australia in which the germination of native species was suppressed in burnt

C. ciliaris-dominated vegetation (Pitts and Albrecht 2000). Reduced perennial grass cover in

burnt plots may have resulted in reduced competition for resources, providing an opportunity

for the establishment and growth of other species. Temporary reductions in cover, rather than

permanent removal, affect species richness. For example, in tallgrass prairie a surge in

herbaceous species richness and frequency was attributed to the removal of shading by the

dominant grasses, rather than plant death since numbers of dominant plants were not reduced

(Copeland et al. 2002). In addition, burning may have directly promoted the germination of

some species, for example legumes. Fire-enhanced pulses of forb abundance have also been

171

reported (Williams et al. 2003b). Measured changes in species richness, percentage cover

and frequency of occurrence may be due to changes in germination patterns and/or changes

in individual plant size (larger plants have a greater chance of being included in a quadrat).

However, the cause of greater abundance of individual species found here remains unclear

since fire effects on germination versus plant growth were not differentiated in this study.

Species richness may be increased by fire if species already in the standing vegetation are

not eliminated and the opening up of the stand and removal of litter allows for the

recruitment of additional species (Daubenmire 1968). Gill (1975) reported that studies had

generally found that species richness either declined or remained unchanged after fire in

Australian communities. However, the effects of fire differ between communities and more

recent work has reported increases in species richness with burning. In wet eucalypt forests

(Bond and van Wilgen 1996), heathlands (refer Cheal 1996) and eucalypt savanna (Williams

et al. 2003b) species richness has been found to increase after fire but decline over time. Fire

has been reported to increase species richness in grasslands in Australia (Morgan 1999). It is

also considered to have a strong effect on species richness in overseas grasslands with both

increases (Walker and Peet 1983; DiTomaso et al. 1999) and decreases in richness being

reported (refer Collins and Gibson 1990). The effects of fire on species richness may vary

depending on the season of burning (Parsons and Stohlgren 1989; Copeland et al. 2002). In

this study grass, forb and legume species richness were increased by burning. However,

season of burning effects were only found for legume species richness at Dalrymple. While it

is likely that this resulted from fire-promoted germination of legumes, it is also possible that

greater legume richness reflects the greater reductions in relative C. ciliaris cover in the early

dry season burn treatment relative to the other treatments.

Measures of species richness are strongly scale-dependent with richness in small plots (1 m2)

often poorly correlated with richness in larger plots within which they are nested (Bond and

van Wilgen 1996). Burning increased species richness at the 1 m2 scale but had no effect on

species richness at the plot scale (400 m2). Others have also reported differences in fire

effects on species richness at different scales. Williams et al. (2003b) found that the species

richness of eucalypt savanna in northern Queensland was unaffected by fire at the 100 m2

scale but increased at the 1 m2 scale. The increase in species richness at the small scale but

not at the large scale suggests an increase in plant density and/or size rather than an increase

in the number of species at the site. This appears to be the case at Dalrymple where fewer

species were found after the fires than before. After the fires at Moorrinya, there was a 41%

increase in the number of species found. However, only about half of these additional

species occurred exclusively in burnt plots. Given that there was no effect on species

172

richness at the plot scale, fire effects on species richness appear to be due more to increases

in plant density and/or size than to the addition of new species at this site also.

The seasonality of fire can influence grassland species composition and biomass (Parsons

and Stohlgren 1989). However, in these C. ciliaris-dominated communities the timing of fire

had relatively minor effects on vegetation responses to burning. Season of burning may be a

major influence on species composition in communities where there are significant

differences in the phenology of plant groups (Howe 1994) since fire will alter the

competitive abilities of plants by causing damage at different stages in their development

(Copeland et al. 2002). However, species in these grasslands generally exhibit similar, strong

seasonal growth patterns. Seed dormancy is broken by heat during the dry season (Mott and

Andrew 1985a; McIvor and Gardener 1994) and most seed germination occurs at the start of

the wet season (McIvor 1987). Above-ground vegetative growth also commences after the

first significant rains of the wet season and perennial plants die back at the end of the wet

season. In these communities, rainfall has an over-riding influence on vegetation dynamics.

The accumulation of biomass after fire is dependent on post-fire moisture availability (Lunt

and Morgan 2002). Post-fire rainfall also has a major effect on the recruitment success of

newly emerged plants (Bond and van Wilgen 1996). Consequently, the amount and timing of

rainfall events may be more significant than the timing of fire in controlling post-fire

development (Morgan 1999).

Given that climate and nutrients, rather than fire, are the key determinants of vegetation

composition and structure in these savannas (Williams et al. 2002) and that there are

relatively minor differences in phenology between the key species, fire-induced changes in

these grasslands will be more subtle than in some other communities. In addition, single fires

are unlikely to cause significant changes in herbaceous community composition. Further

studies are required to determine if C. ciliaris abundance can be decreased by manipulating

the frequency of fire and its timing to exploit differences in C. ciliaris and native species

phenology and/or other attributes.

173

3.11. GENERAL DISCUSSION:

MANAGING CENCHRUS CILIARIS WITH FIRE

3.11.1. THE EFFECTS OF SEASON OF BURNING ON CENCHRUS CILIARIS-

DOMINATED GRASSLAND

In perennial grasslands, one of the major constraints to vegetation change is the persistence

of the established plants. Established plants greatly influence the survival and establishment

of seedlings via competition for light (Copeland et al. 2002), water and nutrients (Cook

1980). Even where resources are made available by the mortality of an individual, they are

often sequestered by the remaining plants (Lauenroth and Aguilera 1998). The removal of

plant competition is a key factor in promoting seedling survival (Cheplick 1998; Cheplick

and Quinn 1988) and is critical to altering the composition of C. ciliaris-dominated

vegetation.

Fire is an important agent in structuring communities since, by removing perennial plant

competition, it creates the potential for vegetation change (Bond and van Wilgen 1996).

Effects of fire on ecosystem components and processes, such as seed banks and flowering,

will be irrelevant in promoting change unless resources are made available to support the

establishment of new individuals. Although the relatively low intensity fires reported here

appeared to have only minor effects on these C. ciliaris-dominated communities, it was

shown that fire could kill C. ciliaris plants and reduce its cover. Fires of higher intensity can

be expected to have greater effects on C. ciliaris abundance and strategic use of fire may

offer opportunities to disadvantage C. ciliaris and promote native species. However, the

long-term responses of these communities to particular fire regimes are difficult to predict

from current knowledge.

The studies reported in the previous sections were conducted to help evaluate fire as a tool to

manipulate the composition of C. ciliaris-dominated vegetation. Overall, the fires had little

impact on herbaceous species composition of these grasslands. Although short-term studies

of single fires are unlikely to offer a definitive recommendation regarding the use of fire to

manage invasive species, the results from these studies provide information that will aid the

development of future work aimed at developing strategies for managing C. ciliaris. The

management implications of the results of these studies and the use of fire as a tool for

managing C. ciliaris-dominated vegetation are discussed below.

174

3.11.1.1. Effects of fire on establishment sites

The fires achieved in these studies probably had little effect on establishment sites. Few C.

ciliaris plants were killed and nutrient levels appeared unaltered. Burning did reduce

perennial grass cover. However, it is unclear how much of this reduction represents

permanent effects on the growth of the existing plants. Temporary reductions in cover are

unlikely to facilitate the establishment of new perennial plants. Importantly, I found that C.

ciliaris is not always promoted by fire, in contrast to the general perception in the invasive

plant literature (Low 1997; Franks et al. 2000; Butler and Fairfax 2003). The positive fire

loop model used to describe C. ciliaris-fire interactions (Butler and Fairfax 2003) was not

supported by the results of this study. Clearly, the C. ciliaris abundance-fire relationship is

complex and a better understanding of the interactions of nutrient and moisture conditions on

this relationship is needed.

Burning effects on establishment sites can be expected to be more significant with higher

intensity fires since plant mortality is often greater (Williams J.R. 1995) and cover removal

more complete (Williams and Cook 2001) with higher intensity fires. The estimated

intensities achieved in my fires (330-3030 kWm-1) were relatively low (section 2.3).

Williams and Cook (2001) reported that early dry season fires in Kadadu National Park

averaged about 2000 kWm-1 while late dry season fires averaged about 8000 kWm-1. Grice

(1997) reported an average intensity of about 7000 kWm-1 for August fires in shrub-invaded

grassland near Townsville and Williams et al. (2003b) reported intensities of around 1500

kWm-1 for early dry season fires and 5500 kWm-1 for late dry season fires in coastal savanna

near Townsville, northern Queensland.

Although there was some evidence that burning effects were related to fire intensity, I was

unable to unequivocally demonstrate vegetation responses to fire intensity. The range of plot

fire intensities achieved was relatively small and the uneven spread of intensities of

Dalrymple plots prevented valid assessment of any relationships. It is also important to point

out that the estimates, which are based on fuel loads and fire speeds, were approximate. An

average site fuel load was used. However, biomass levels varied within the plots. Only one

fire speed was recorded per plot, although fire speeds clearly changed across each plot. As

pointed out by Gill and Knight (1991), moving fires present difficulties in measurement

because they pulsate and surge. More important perhaps is the fact that fire intensity may or

may not be related to the intensity of disturbance (Christensen 1985). That is, average plot

fire intensities may not adequately describe the local fire conditions that are critical in

affecting individual plants (section 3.2.3.1). Difficulties associated with measuring fire are

175

well recognized and Lonsdale and Braithwaite (1991) noted that it was not possible to

measure the most important variables at the precise time of ignition.

Although clear relationships between fire intensity and vegetation response were not found,

it is likely that relatively high intensity fires will be needed to eliminate or reduce C. ciliaris

competition. Fires producing higher temperatures are likely to result in higher plant

mortality. Alexander and D’Antonio (2003) noted that the success of fire as a control

treatment for any exotic species is likely to be strongly dependent on fire temperature. An

important point demonstrated by my study is that season of burning may not be a good

predictor of fire intensity. Although late dry season fires are generally more intense than

early dry season fires (Williams et al. 1997), others have also noted that season of burning is

not necessarily a good predictor of fire intensity, and, in some years relatively intense fires

can occur in the early dry season (Williams J. R. 1995). Consequently, prevailing weather

conditions, rather than time of year, may be more important in determining the magnitude of

the fire effect in these grasslands. It is also important to appreciate that opportunities for

achieving high intensity, high mortality fires may be relatively rare since the use of fire is

limited by fuel availability (Hitchmough et al. 1994). The imposition of high intensity fires

will be constrained by seasonal and grazing conditions with opportunities for effective fires

being limited to good seasons when fuel levels are high.

It may be easier to achieve fatal temperatures in C. ciliaris patches than in patches of native

species. The amount of biomass produced by individual species influences the fire intensities

they experience. For example, it has been suggested that Chrysopogon fallax is relatively

unaffected by fire because it grows in relatively small clumps and does not accumulate large

quantities of litter around it, so that immediate heat loads are low (Smith 1960). Compared

with native species, C. ciliaris may be more susceptible to medium intensity fires since it

produces more biomass than native species (Humphries et al. 1991; Latz 1991), creating its

own higher fuel loads. The thick litter accumulation observed in some C. ciliaris patches

(section 3.6) would enhance fuel levels. However, C. ciliaris biomass is not always greater

than that of native species (section 2.2).

The implications of high intensity fires on plant mortality in these systems remain to be

investigated. In this study it was not possible to monitor the effects of fire on native species

in any detail. Although the resident perennial grasses share a similar morphology and

phenology and all have evolved with exposure to fire and survive burning (McIvor and Orr

1991), subtle differences can influence responses to fire. For example, the relatively exposed

growing points of Themeda triandra and shallow roots of Sarga (Sorghum) plumosum are

176

believed to make these species more vulnerable to fire than deep rooted species such as C.

fallax (Smith 1960). In contrast to work by Pitt and Albrecht (2000), the increase in small-

scale species richness found with burning (section 3.10) indicates that at least some native

species respond positively to burning C. ciliaris-dominated vegetation. It is unclear how the

resident perennial species in these grasslands will respond, relative to each other, to high

intensity fires.

As well as removing established plants, fire may affect establishment site availability via

effects on soil nutrients and surface cover. Although no changes in nutrient levels with

burning were detected, more intense fires may affect nutrient availability. Changes in

nutrient availability have the potential to alter competitive interactions (Howden 1988).

Cenchrus ciliaris is favoured by moderate to high nutrient conditions (Humphries 1967;

McIvor 1984) and post-fire flushes of nutrients may give it an advantage over less nutrient-

demanding native species. Litter can also be very important in determining species

composition (Lane and Williams 1997). My hypothesis that the thick litter mats that

sometimes form in between C. ciliaris plants would inhibit seedling emergence was not

supported. Matted litter was found to benefit C. ciliaris emergence (section 3.6.3). Overall,

litter cover reduced C. ciliaris and Heteropogon contortus emergence compared with bare

soil under ‘controlled’ conditions. However, in the field the effects of litter are likely to be

more favourable for seedling emergence. It is not known how the extensive removal of litter

by high intensity fires will affect the seedling recruitment patterns of key species.

3.11.1.2. Effects of fire on propagule supply

Burning appears to have little effect on propagule supply in these grasslands, although the

low numbers of perennial grasses detected in the soil seed banks prevented any

determination of fire effects on perennial grass seed bank composition (section 3.7). The low

numbers of C. ciliaris and native perennial grass seed found in the seed banks suggest that

these species may not be present to establish when sites become available (McIvor 1987;

refer McIvor and Gardener 1994). This may present an opportunity to manipulate post-fire

seedling recruitment. Sowing desirable native species may be effective in promoting species

change since there may be little competition from naturally sown seedlings. However, the

seed bank study results should be interpreted with caution. It is possible that the seed bank

composition determined by the germination method (section 3.7.2) does not accurately

reflect the actual composition and it was not possible to identify the different species of grass

seedlings that emerged in the field (section 3.9). Longer-term studies are required to

177

investigate seedling recruitment dynamics in these grasslands.

Some changes in relative abundance of species in seedling populations following fire may be

explained in terms of variations in diaspore morphology (Peart 1984). Differences in

diaspore morphology were expected to result in differences in abundance in the post-fire

seed banks and seedling populations. The diaspore of C. ciliaris appears vulnerable to fire

since it has no effective seed burial mechanism, remains on or close to the soil surface

(Hacker 1989) and is readily destroyed or damaged by fire (Ernst 1991). In contrast, some

native species such as T. triandra and H. contortus, have seeds with twisting hydroscopic

awns that help bury them into the soil, thus protecting them from destruction by fire (Dyer et

al. 1997). In my study, too few perennial grass plants emerged from the soil seed banks to

determine any patterns relating to season of burning and I was unable to identify grass

seedlings to species level in the field. The results from the litter study (section 3.6.) showed

that H. contortus may be more vulnerable to fire than expected if seed falls on litter.

However, the importance of diaspore type in relation to vulnerability to fire in these

grasslands is unclear from this work.

Flowering patterns, rather than diaspore type, may be more important in influencing fire

effects in these grasslands. Cenchrus ciliaris flowering did not appear to be promoted by

burning, contrary to what was expected from other accounts (refer Humphries et al. 1991).

Seed head production in C. ciliaris in central Australia is determined by rainfall (Bosch and

Dudzinski 1984) and it is likely that rainfall, rather than fire, is the major influence on

flowering in C. ciliaris in the communities studied here.

The timing of fire in relation to flowering can have significant effects on grassland

composition (Howe 1994). The perennial grasses present in these C. ciliaris-dominated

grasslands vary in the details of their flowering responses. For example, in its northerly

range Heteropogon contortus flowering is controlled by day length and plants flower at the

end of summer (Orr 1998). In contrast, flowering in Astrebla spp. can occur throughout the

year and seed production is promoted by rainfall and fire (Scanlan 1980). Cenchrus ciliaris

also produces flower heads throughout the year (Hacker 1989). By adding seed slowly over

the season, the seed inputs of species such as C. ciliaris may be less affected by fire. Fires at

specific times may significantly affect seed inputs of species such as H. contortus that

produce all their seed at one time. A year’s H. contortus seed may be drastically reduced by

early dry season burning since the seed may still be held in seed heads or be on the soil

surface (Walker et al. 1983). Further work is required to determine how season of burning

affects the inputs of viable seed of C. ciliaris and native species in these grasslands.

178

3.11.2. FIRE AS A MANAGEMENT TOOL – ARE THERE OPPORTUNITIES TO

REDUCE CENCHRUS CILIARIS ABUNDANCE?

Fire is an effective tool for managing both woody and herbaceous invasive species. In

natural and semi-natural systems it is the cheapest tool available (Christensen and Burrows

1986) and has the advantage that it can be easily used over large areas (Grice et al. 2000).

However, fire may be ineffective in some communities and may be inappropriate for certain

species. It may be highly effective for manipulating vegetation composition in communities

where there are significant differences in fire tolerance and/or phenology between plant

groups (Howe 1994) or where fire plays a major role in ecosystem processes (Williams et al.

2002). Communities vary in these characteristics and, consequently, in their responsiveness

to fire. In the C. ciliaris-dominated grasslands studied here, the effects of fire may be limited

since differences in phenology and morphology between C. ciliaris and the native species

are relatively subtle and climate, rather than fire, is a dominant force influencing vegetation

dynamics (refer Williams et al. 2002). The value of fire as a management tool also depends

on the invasive species involved. Recent literature on fire and exotic species suggests that

fire generally tends to promote rather than discourage introduced species (D’Antonio 2000;

Grice et al. 2000; Wilson and Mudita 2000)). Nevertheless, there is evidence that many

species can be controlled by fire (D’Antonio 2000) and despite its limitations, it is a useful

tool for decreasing the vigour of species that may otherwise dominate and exclude other

species (Stuwe and Parsons 1977).

Fire may result in compositional changes where there are differences in plant condition (size

and health), morphology and/or differences in life strategy. Although differences between C.

ciliaris and associated native perennial grass species are relatively subtle compared with

those found between co-occurring species in some other habitats, there are differences that

can perhaps be exploited. Differences in phenology may provide opportunities to reduce C.

ciliaris abundance using fire while having less effect on the native species. Cenchrus ciliaris

is reported to remain green longer and cure later than native species (Cavaya 1988;

Humphries et al. 1991). This may make it more susceptible to early dry season burns than

native species that have senesced earlier. The higher C. ciliaris mortality with early dry

season burns at Dalrymple may reflect this vulnerability (section 3.4.4). Cenchrus ciliaris is

also noted for its ability to respond quickly to rainfall (Latz 1991). Therefore, opportunistic

fires following out-of-season rainfall events when the native vegetation is dormant may also

result in increased C. ciliaris mortality.

Although season of burning had only minor effects in this study, strategic timing of fire may

179

be vital in manipulating C. ciliaris abundance. Dry season fires were investigated here.

However, it has been noted elsewhere that burning in the dry season when grasses have

seeded and become dormant generally does not affect the composition of perennial

grasslands (Mott and Andrew 1985a). In contrast, wet season burning can result in changes

in yield and composition (Smith 1960). The fast response of C. ciliaris to rainfall, noted

above, may result in significant season of burning effects. Early wet season burning may kill

C. ciliaris plants while slower responding native species may be unaffected. Burning after

substantial rainfall in the dry season may also selectively kill C. ciliaris. However, the

application of opportunistic burning requires considerable resources. Also, as was the case in

this work, sites may not be accessible at times during the wet season.

The apparent lack of C. ciliaris and perennial grass seed in the soil seed banks presents

opportunities to manipulate post-fire recruitment. Since so little perennial grass seed appears

available for establishment in fire-created micro-sites, sowing of native perennial seed may

be an effective means of shifting vegetation composition. Apart from low numbers of C.

ciliaris seedlings, due to low seed availability, it also appears that C. ciliaris seedlings may

be competitively inferior to some native species. Although it is considered a highly

competitive species, the slower emergence of C. ciliaris seedlings compared with H.

contortus seedlings observed in the litter cover study (section 3.6) suggests that C. ciliaris

may be at a competitive disadvantage at the seedling stage. The order of arrival of species

can be important in determining which species dominate particular sites (refer D’Antonio et

al. 2001). Seedlings that germinate first may gain a competitive advantage (section 3.9.4)

and the faster emergence of H. contortus may give it a head start in sequestering resources,

enabling it to out-compete C. ciliaris for the occupancy of new sites. Seedling interactions in

these grasslands are of considerable interest and I am unaware of any studies investigating

competitive interactions between C. ciliaris and native seedlings.

Determining the effects of fire on these systems is problematic and the work reported here

does not allow a comprehensive assessment of the efficacy of fire as a tool for manipulating

C. ciliaris-dominated grasslands. The studies were limited by a number of constraints not

uncommon to ecological studies. Low levels of replication limit the statistical power to

detect treatment effects in these innately variable systems. The time frame of

experimentation and monitoring was short and access to field sites was hindered by weather

and lack of resources. Added to these logistic constraints is the complexity of the systems

being studied. The dynamics of savanna systems, where vegetation change is contingent

upon interactions between fire, weather conditions, soil conditions, competition, herbivores

and other disturbances are notoriously complex (Walker et al. 1981; refer Bond and van

180

Wilgen 1996). No two fires or the conditions in which they occur are alike and both pre- and

post-fire conditions are unique and tend to produce differences in results (Vogl 1974;

D’Antonio 2000). Noble and Grice (2002) noted that, due to the unpredictability of rainfall,

there was rarely any ordered succession in plant communities following fire in semi arid and

tropical grasslands. In addition, the effects of fire may be less important in these grasslands

than in other ecosystems where critical ecosystem processes such as seed germination are

controlled by fire (Lunt and Morgan 2002). Consequently, fire-induced changes may be

more subtle than in other communities. The dynamics of these C. ciliaris-dominated

communities are chiefly climate driven. Responses to burning are largely determined by pre-

and post-fire weather conditions making both predictions (Walker et al. 1981) and

interpretation (Norman 1969) of experimental results difficult.

3.11.3. CONCLUSIONS

There is an urgent need for management strategies that reduce, prevent or contain invasive

plant invasions (Adair and Groves 1998; Wilson and Mudita 2000). While there is a large

literature on biological invasions and invasion processes, there appears to be less published

information specifically relating to the control of environmental weeds (Barrow 1995). In the

case of C. ciliaris, there is a vast literature dealing with this species in an agricultural context

as a desirable pasture species. More recently, the negative effects of C. ciliaris have been

highlighted. However, there is little published information specifically relating to the

management of C. ciliaris in situations where it is not wanted (see Pitts and Albrecht 2000).

Fire plays a critical role in structuring many plant communities (refer Tyler 1995) and is

widely used for managing ecosystems worldwide (Bond and van Wilgen 1996). However, its

usefulness in managing invasive grasses is limited. Fire may promote invasive grasses

(Friday et al. 1999; Rossiter et al. 2003) and where fire has been found to reduce exotic grass

abundance, increases in non-target exotics may occur (Parsons and Stohlgren 1989).

Consequently, the use of fire for invasive plant control requires caution and plant and fire

factors must be critically evaluated (D’Antonio 2000).

At the beginning of this chapter I asked if fire could be used to reduce the abundance of C.

ciliaris and promote the recruitment of native species? The removal of established C. ciliaris

plants is key to promoting change in these grasslands and fire may be useful in promoting

181

compositional change if it removes plants and reduces C. ciliaris competition. In this study,

C. ciliaris abundance was unaffected or reduced by fire, at least temporarily, and burning

resulted in increased abundance of some native species. The intensity of fire is likely to be a

major factor determining its effectiveness in removing C. ciliaris. However, the timing of

fire may also be important. Relatively subtle differences in growth and phenology between

C. ciliaris and the native species in these grasslands may be opportunistically exploited to

disadvantage C. ciliaris. However opportunities to achieve effective fires will be limited by

fuel availability and fires alone are unlikely to result in significant shifts in composition.

Fire is a secondary determinant of vegetation composition in these systems and its effects are

greatly modified by climatic conditions. However, since there are few cost-effective

management options available for managing vegetation at large scales (Noble and Grice

2002) it is necessary to explore any possibilities that fire may offer. The fires reported here

had no major effects on the composition of C. ciliaris-dominated grassland. It is important to

distinguish between shorter-term changes occurring after a single fire event and the longer-

term effects resulting from a particular fire regime (Raison 1979). As pointed out by Lunt

and Morgan (2002), short-term impacts of single fires provide little insight into the effects of

consecutive fires. Longer-term studies are required to more fully understand fire-vegetation-

climate interactions in these grasslands. The effects of high intensity fires on these

grasslands also need to be determined if strategies for containing or reducing the spread of C.

ciliaris are to be developed.

182

CHAPTER 4. CENCHRUS CILIARIS AS AN INVASIVE SPECIES –

FUTURE RESEARCH QUESTIONS

Cenchrus ciliaris is perhaps one of Australia’s most controversial imports. From a pastoral

point of view, this now extensively distributed grass is highly valued for its perseverance

and productivity under harsh conditions (Paull and Lee 1978). The success of C. ciliaris as a

pasture species in Australia’s arid and semi-arid zones is due to traits such as its ease of

establishment, rapid growth, adaptation to a range of soils and its tolerance of drought and

grazing (Franks 2002). However, these traits have led to its spread into non-target areas with

reported negative impacts on native species and fire regimes (Humphries et al. 1991; Low

1997). From a conservation/sustainable land management point of view, C. ciliaris is

considered to be one of Australia’s worst environmental weeds (Humphries et al. 1991;

State of the Environment 1996).

Despite these conflicting perceptions, there are relatively few published data quantifying the

environmental effects of C. ciliaris. Studies investigating this species in an environmental

conservation context have been conducted in central Australia, where this species is

considered to be a major threat (Pitts and Albrecht 2000; Best 1998), and in central

Queensland (Fairfax and Fensham 2000; Franks 2002; Butler and Fairfax 2003). The study

reported here is, to my knowledge, the first specifically investigating C. ciliaris from an

environmental perspective in north Queensland. I addressed two issues: (1) the impacts of

C. ciliaris on herbaceous species richness and (2) the use of fire to manage C. ciliaris.

Quantitative data describing the effects of C. ciliaris on biodiversity are critical to achieving

rational debate about the values and risks associated with this species. My study and others

(Fairfax and Fensham 2000; Franks 2002) provide correlative data showing that C. ciliaris

is associated with reduced herbaceous species richness. However, these studies fall short of

demonstrating a cause and effect relationship. It is difficult to determine the role of invasive

plants in reducing biodiversity since plant invasion often occurs simultaneously with other

disturbances (Vitousek 1990). Species richness is influenced by factors such as drought and

grazing, the impacts of which are confounded with those of the invading species.

Consequently, obtaining evidence of a causal relationship between C. ciliaris and species

richness is a considerable challenge. In the Hillgrove study (section 2.2), I was able to

demonstrate a relationship between C. ciliaris biomass and herbaceous species richness.

However, the cause of this association was not determined. Manipulative studies are

required to identify the mechanisms underlying C. ciliaris-species richness relationships.

Cenchrus ciliaris is considered to be a successful competitor (McIvor 2003) and it is also

183

suggested that it has allelopathic effects (Cheam 1984ab; Nurdin and Fulbright 1990). These

traits need to be investigated to help understand C. ciliaris-biodiversity relationships.

Both spatial and temporal scales have important implications for C. ciliaris-biodiversity

relationships. Although C. ciliaris has been associated with reduced herbaceous species

richness at small scales, it is not clear how dominance over large areas affects the survival

of native species. Relationships between C. ciliaris and native species may also change over

time as C. ciliaris invasion alters ecosystem processes, such as fire regimes, and changes in

ecosystem processes affect C. ciliaris dominance. Studies investigating individual species

abundance patterns in relation to C. ciliaris abundance are needed to determine whether

native species can persist in C. ciliaris-dominated communities, or whether C. ciliaris

invasion leads to loss of species over time.

There is an urgent need for effective control strategies for C. ciliaris that have minimal

negative environmental impacts (Pitts and Albrecht 2000). Fire is one of the few tools

available for managing vegetation on an extensive scale, and, despite its limitations

(D’Antonio 2000), manipulation of the fire regime has been suggested as a method to

maintain or restore C. ciliaris-invaded grasslands overseas (Daehler and Carino 1998). In

Australia, some workers have discounted fire as a useful, long-term solution to C. ciliaris

invasion since they found that C. ciliaris recovered rapidly and completely after fire (Pitts

and Albrecht 2000). However, the results from my short-term fire study demonstrate that C.

ciliaris-fire interactions are more complex than is suggested from the positive fire loop

model proposed (Butler and Fairfax 2003). I found that fire killed C. ciliaris plants and

reduced its cover and that this was associated with an increase in native species abundance.

Post-fire growing conditions, particularly rainfall, influence vegetation responses to burning

and responses will differ across sites, depending on factors such as soil nutrient status, soil

seed bank composition and grazing regime. The fires in this study were generally patchy

and of low intensity. Of interest is how these grasslands respond to more extensive, higher

intensity fires and the implications of site conditions on these responses.

Although season of burning had relatively minor effects on these grasslands, the timing of

fire can be critical in determining outcomes (Daubemire 1968). More intense fires are likely

to be more effective in reducing C. ciliaris biomass. However, as shown here, weather

conditions, rather than season, may be more important for determining fire intensity in these

systems. Studies are also required to investigate the effects of wet season fires. Burning at

the start of the wet season or after rainfall in the dry season may selectively kill C. ciliaris,

given its fast initiation of growth following rainfall.

184

Manipulation of the fire regime may offer opportunities to disadvantage C. ciliaris relative

to native species. However, it is unlikely that fire alone will result in significant shifts in

composition. Additional strategies are required. The slow emergence of C. ciliaris seedlings

compared with H. contortus seedlings observed in the litter study (section 3.6), together

with the probability of relatively low numbers of perennial grass seed in the seed banks of

these communities (section 3.7), suggests that over-sowing with native species may be

beneficial.

In summary, a number of research questions have been raised in this thesis. Future studies

are required to:

• identify the mechanism (s) responsible for negative associations between C. ciliaris

and herbaceous species richness;

• determine the impacts of C. ciliaris invasion on the survival of native species;

• further investigate season of burning effects to determine strategies for reducing C.

ciliaris abundance (can physiological differences be exploited?);

• determine the impacts of high intensity fires in these communities (how do native

species respond to high intensity fires? What are the implications for nutrient/soil

stability?);

• develop a better understanding of plant community-fire-moisture regime (rainfall)

interactions, particularly in relation to the effects of burning on perennial grass seed

inputs and recruitment patterns; and

• investigate the competitive ability of C. ciliaris seedlings. Is over-sowing with

native species an option?

185

REFERENCES Achhireddy N. R. and Singh M. (1984) Allelopathic effects of lantana (Lantana camara) on Milkweed vine (Morrenia odorata). Weed Science 32, 757-761.

Adair R. (1995) The threat of environmental weeds to biodiversity in Australia: a search for solutions. In: Conserving Biodiversity: Threats and Solutions (eds R. A. Bradstock, T. D. Auld, D. A. Keith, R. T. Kingsford, D. Lunney and D. P. Sivertsen) pp. 184-201. Surrey Beatty and Sons, Chipping Norton, NSW, Australia.

Adair R. J. and Groves R. H. (1998) Impact of Environmental Weeds on Biodiversity: A Review and Development of a Methodology. Biodiversity Group, Environment Australia, Canberra.

Al-Mufti M. M., Sydes C. L., Furness S. B., Grime J. P. and Band S. R. (1977) A quantitative analysis of shoot phenology and dominance in herbaceous vegetation. Journal of Ecology 65, 759-791.

Albrecht D. and Pitts B. ( 1997) Buffel grass on Parks and Wildlife Commission Reserves. Alice Springs Rural Review 47, 7.

Aldous A. E. (1934) cited in Daubenmire R. E. (1968), Ecology of fire in grasslands. Advances in Ecological Research 5, 209-266.

Alexander J. M. and D'Antonio C. M. (2003) Seed bank dynamics of french broom in coastal California Grasslands: effects of stand age and prescribed burning on control and restoration. Restoration Ecology 11, 185-197.

Anderson A. N., Braithwaite R. W., Cook G. D., Corbett L. K., Williams R. J., Douglas M. M., Gill A. M., Setterfield S. A. and Muller W. J. (1998) Fire research for conservation management in tropical savannas: Introducing the Kapalga fire experiment. Australian Journal of Ecology 23, 95-110.

Anderson E. R. (1972) Emergence of buffel grass (Cenchrus ciliaris) from seed after flooding. Queensland Journal of Agricultural and Animal Sciences 29, 167-172.

Andrew M. H. and Mott J. J. (1983) Annuals with transient seed banks: the population biology of indigenous Sorghum species of tropical north-west Australia. Australian Journal of Ecology 8, 265-276.

Ash A. J., Corfield J. and Ksiksi T. (2002) The Ecograze Project: Developing Guidelines to Better Manage Grazing Country. CSIRO Sustainable Ecosystems, Townsville, Australia.

Asher R. (1991) cited in Daehler and Strong (1996), Status, prediction and prevention of introduced cordgrass Spartina spp. invasions in Pacific estuaries, USA. Biological Conservation 78, 51-58.

Auld T. D. (1986) Population dynamics of the shrub Acacia suaveolens (Sm.) Willd.: fire and the transition to seedlings. Australian Journal of Ecology 11, 373-385.

Auld T. D. and Bradstock R. A. (1996) Soil temperatures after the passage of a fire: do they influence the germination of buried seed? Australian Journal of Ecology 21, 106-109.

Auld T. D. and O'Connell M. A. (1991) Predicting patterns of post-fire germination in 35

186

eastern Australian Fabaceae. Australian Journal of Ecology 16, 53-70.

Babu R. C. and Kandasamy O. S. (1997) Allelopathic effects of Eucalyptus globulus Labill. on Cyperus rotundus L. and Cynodon dactylon L. Pers. Journal of Agronomy and Crop Science 179, 123-126.

Back P. V. (2001) Regrowth control and its effects on buffel grass pastures. In: Buffel Grass Symposium. Proceedings of Workshop Held at Theodore on 21-23 February 2000 (eds J. Kyte, H. Lees, I. Loxton and B. Cook) pp. 21-22. Department of Primary Industries, Brisbane, Australia.

Bailey F. M. (1906) The Weeds and Suspected Poisonous Plants of Queensland. Goverment Printer, Brisbane.

Baird A. M. (1977) Regeneration after fire in King's Park, Perth, Western Australia. Journal of the Royal Society of Western Australia 60, 1-22.

Barbour M. G. and Johnson A. F. (1977) cited in Woods (1997), Community response to plant invasion. In: Assessment and Management of Plant Invasions (eds J. O. Luken and J. W. Thieret) pp. 56-68. Springer-Verlag, New York.

Barritt A. R. and Facelli J. M. (2001) Effects of Casuarina pauper litter and grove soil on emergence and growth of understorey species in arid lands of South Australia. Journal of Arid Environments 49, 569-579.

Barrow P. (1995) The ecology and management of Gamba grass (Andropogon gayanus Kunth.). Northern Territory Department of Primary Industry and Fisheries, Northern Territory.

Batianoff G. N. and Franks A. J. (1998) Weed invasions of the tropical Mackay coast, Queensland, Australia. Plant Protection Quarterly 13, 123-130.

Beatley J. C. (1966) Ecological status of introduced brome grass (Bromus spp.) in desert vegetation of southern Nevada. Ecology 47, 548-554.

Bebawi F. F., Campbell S. D., Lindsay A. M. and Grice A. G. (2000) Impact of fire on rubber vine (Cryptostegia grandiflora R.Br.) and associated pasture and germinable seed bank in a sub-riparian habitat of north Queensland. Plant Protection Quarterly 15, 62-66.

Beerling D. J. (1995) General aspects of plant invasions: an overview. In: Plant Invasions: General Aspects and Special Problems (ed. P. Pysek) pp. 237-247. SPB Academic Publishing, Amsterdam.

Begon M., Harper J. L. and Townsend C. R. (1990) Ecology: Individuals, Populations and Communities. Blackwell Scientific Publications, London.

Belcher J. W. and Wilson S. D. (1989) Leafy spurge and the species composition of a mixed-grass prairie. Journal of Range Management 42, 172-175.

Bell D. T. (1999) The process of germination in Australian species. Australian Journal of Botany 47, 475-517.

Belnap J. and Phillips S. L. (2001) Soil biota in an ungrazed grassland: response to annual grass (Bromus tectorum) invasion. Ecological Applications 11, 1261-1275.

Bergelson J. (1990) Life after death: site pre-emption by the remains of Poa annua. Ecology

187

71, 2157-2165.

Best R. A. (1998) The Effects of Introduced Buffel Grass (Cenchrus ciliaris L., Poaceae) on the Diversity and Abundance of Invertebrates in Semi-Arid Central Australia. Thesis, School of Biological and Environmental Sciences, Faculty of Science, Northern Territory University.

Bigwood D. W. and Inouye D. W. (1988) Spatial pattern analysis of seed banks: an improved method and optimized sampling. Ecology 69, 497-507.

Billings W. D. (1990) Bromus tectorum, a biotic cause of ecosystem impoverishment in the Great Basin. In: The Earth in Transition. Patterns and Processes of Biotic Impoverishment (ed. G. M. Woodwell) pp. 69-80. Cambridge University Press, Cambridge, New York.

Bond W. J. and van Wilgen B. W. (1996) Fire and Plants. Chapman and Hall, London.

Bosch O. J. H. and Dudzinski M. L. (1984) Defoliation and its effects on Enneapogon avenaceus and Cenchrus ciliaris populations during two summer growth periods in central Australian Rangelands. Australian Rangeland Journal 6, 17-25.

Bowman D. M. J. S., Wilson B. A. and Hooper R. J. (1988) Response of Eucalyptus forest and woodland to four fire regimes at Munmalrary, Northern Territory, Australia. Journal of Ecology 76, 215-232.

Boylen C. W., Eichler L. W. and Madsen J. D. (1999) Loss of native aquatic plant species in a community dominated by Eurasian watermilfoil. Hydrobiologia 415, 207-211.

Braithwaite R. W. (1995) cited in Dyer R., Craig A. and Grice A. (1997), Fire in northern pastoral lands. In: Fire in the Management of Northern Australian Pastoral Lands pp. 24-40. The Tropical Grassland Society of Australia, Brisbane, Australia.

Braithwaite R. W., Lonsdale W. M. and Estbergs J. A. (1989) Alien vegetation and native biota in tropical Australia: the impact of Mimosa pigra. Biological Conservation 48, 189-210.

Breytenbach G. J. (1986) cited in Walker L. R. and Smith S. D. (1997), Impacts of invasive plants on community and ecosystem properties. In: Assessment and Management of Plant Invasions (eds J. O. Luken and J. W. Thieret) pp. 69-86. Springer-Verlag, New York.

Bridge B. J., Mott J. J. and Hartigan R. J. (1983) The formation of degraded areas in the dry savanna woodlands of northern Australia. Australian Journal of Soil Research 21, 91-104.

Bridgewater P. B. and Backshall D. J. (1981) Dynamics of some Western Australian ligneous formations with species reference to the invasion of exotic species. Vegetatio 46, 141-148.

Brown C. S. and Rice K. J. (2000) The mark of Zorro: effects of the exotic annual grass Vulpia myuros on California native perennial grasses. Restoration Ecology 8, 10-17.

Brown R. (1985) The effect of severe defoliation on the subsequent growth and development of five rangeland pasture grasses of south-western Queensland. Australian Journal of Ecology 10, 335-343.

Brown S. A. and Ramsay C. M. (1999) Invasion of the Townsville Town Common Environmental Park by the environmental weeds Brachiaria mutica (Para grass) and Panicum maximum (Guinea grass): implications for its future management. Tropical

188

Landscape Ecology and Management Monographs 5, 1-31.

Butler D. W. and Fairfax R. J. (2003) Buffel grass and fire in a gidgee and brigalow woodland: a case study from central Queensland. Ecological Management and Restoration 4, 120-125.

Byram G. M. (1948) cited in Daubenmire R. E. (1968), Ecology of fire in grasslands. Advances in Ecological Research 5, 209-266.

Byram G. M. (1959) Combustion of forest fuels. In: Forest Fire. Control and Use (ed. K. P. Davis) pp. 61-89. McGraw-Hill Book Company, Inc., New York.

Calvo L., Herrero B. and Bravo F. (1999) The influence of fire on the seed bank in the soil of a Quercus faginea forest (NW Spain). Annals of Forest Science 56, 249-258.

Campbell S. D. (1995) cited in Orr, D. M. (1998), A life cycle approach to the population ecology of two tropical grasses in Queensland, Australia. In: Population Biology of Grasses (ed. G. P. Cheplick) pp. 366-389. Cambridge University Press, Cambridge, United Kingdom.

Campbell S. D. and Setter C. L. (2002) Mortality of Prosopis pallida (mesquite) following burning. Australian Journal of Experimental Agriculture 42, 581-586.

Carr G. (1988) Perspectives on environmental weeds in Victoria. In: Weeds on Public Land - an Action Plan for Today. Proceedings of a Symposium Presented by the Weed Science Society of Victoria Inc. and the School of Environmental Science, Monash, Held at Monash University, Clayton Victoria, Australia, 19-20 May, 1988, pp. 52-58. Weed Science Society of Victoria Inc., Victoria, Australia.

Cavaye J. M. (1988) Buffel grass basics. Queensland Agricultural Journal 114, 69-75.

Cavaye J. M. (1991) The Buffel Book. A Guide to Buffel Grass Pasture Development in Queensland. Queensland Department of Primary Industries, Brisbane, Australia.

Chaudhry M. S., Khan M. A. and Nasim F. H. (1999) Effect of light and two isomers of nitrophenyl indole-2-carboxylic acid on germination of Cenchrus ciliaris seeds. Pakistan Journal of Botany 31, 265-270.

Cheal D. (1996) Fire succession in heathlands and implications for vegetation management. In: Fire and Biodiversity. The Effects and Effectiveness of Fire Management. Proceedings of the Conference Held 8-9 October, Footscray, Melbourne., Department of the Environment, Sport and Territories, Biodiversity Unit, Canberra, ACT.

Cheam A. H. (1984a) Allelopathy in buffel grass (Cenchrus ciliaris L.) Part 1. Influence of buffel grass association on calotrope (Calotropis procera (Ait) W.T.Ait.). Australian Weeds 3, 133-136.

Cheam A. H. (1984b) Allelopathy in buffel grass (Cenchrus ciliaris L.) Part II. Site of release and distribution of allelochemical in the soil profile. Australian Weeds 3, 137-139.

Cheney P. and Sullivan A. (1997) Grass Fires, Fuel, Weather and Fire Behaviour. CSIRO Publishing, Melbourne.

Cheplick G. P. (1998) Seed dispersal and seedling establishment in grass populations. In: Population Biology of Grasses (ed. G. P. Cheplick) pp. 84-105. Cambridge University Press, Cambridge, United Kingdom.

189

Cheplick G. P. and Quinn J. A. (1988) Subterranean seed production and population responses to fire in Amphicarpum purshii (Gramineae). Journal of Ecology 76, 263-273.

Christensen N. L. (1985) Shrubland fire regimes and their evolutionary consequences. In: The Ecology of Natural Disturbance and Patch Dynamics (eds S. T. A. Pickett and P. S. White) pp. 85-100. Academic Press, New York, USA.

Christensen N. L. and Abbott I. (1989) Impacts of fire in the eucalypt forest ecosystem of southern Western Australia: a critical review. Australian Forestry 52, 103-121.

Christensen N. L. and Muller C. H. (1975) Effects of fire on factors controlling plant growth in Adenstoma chaparral. Ecological Monographs 45, 29-55.

Christensen P. E. and Burrows N. (1986) Fire: an old tool with a new use. In: Ecology of Biological Invasions: an Australian Perspective (eds R. H. Groves and J. J. Burden) pp. 97-105. Australian Academy of Science, Canberra.

Clifford H. T., Rogers R. W. and Kahler C. (1995) Seasonal variation in the germinable seed banks of southern Queensland eucalypt forest. Proceedings of the Royal Society of Australia 105, 1-5.

Collins S. L. (1990) Introduction: fire as a natural disturbance in tallgrass prairie ecosystems. In: Fire in North American Tallgrass Prairies (eds S. L. Collins and L. L. Wallace) pp. 3-7. University of Oklahoma Press, Norman.

Collins S. L. and Gibson D. J. (1990) Effects of fire on community structure in tallgrass and mixed prairie. In: Fire in North American Tallgrass Prairies (eds S. L. Collins and L. L. Wallace) pp. 81-98. University of Oklahoma Press, Norman.

Cook G. D. (1992) The effects of fire on nutrient losses from top end savannas. In: Conservation and Development Issues in North Australia (eds I. Moffatt and A. Webb) pp. 123-129. North Australia Research Unit, Darwin.

Cook G. D. (1994) The fate of nutrients during fires in a tropical savanna. Australian Journal of Ecology 19, 359-365.

Cook G. D. (2001) Whats in the air? In: Savanna Burning. Understanding and Using Fire in Northern Australia (eds R. Dyer, P. Jacklyn, I. Partridge, J. Russell-Smith and D. Williams) pp. 48. Tropical Savanna CRC, Darwin.

Cook R. E. (1979) Patterns of juvenile mortality and recruitment in plants. In: Topics in Plant Population Biology (eds O. T. Solbrig, S. Jain, G. Johnson and P. H. Raven) pp. 207-231. Columbia University Press, New York.

Cook S. J. (1975) Germination of surface-sown buffel grass seed. In: Australian Conference on Tropical Pastures, James Cook University of North Queensland, Townsville, Queensland, pp. 4-56 to 4-60. CSIRO, Canberra.

Cook S. J. (1980) Establishing pasture species in existing swards: a review. Tropical Grasslands 14, p181-187.

Cook S. J. (1984) Establishment of four pasture grasses and Siratro from seed oversown into dense and open speargrass pastures. Australian Journal of Experimental Agriculture and Animal Husbandry 24, 360-369.

Cook S. J., Clem R. L., MacLeod N. D. and Walsh P. A. (1993b) Tropical pasture

190

establishment. 7. Sowing methods for pasture establishment in northern Australia. Tropical Grasslands 27, 335-343.

Cook S. J., Gilbert M. A. and Shelton H. M. (1993a) Tropical pasture establishment. 3. Impact of plant competition on seedling growth and survival. Tropical Grasslands 27, 291-301.

Copeland T. E., Sluis W. and Howe H. F. (2002) Fire season and dominance in an Illinois tallgrass prairie restoration. Restoration Ecology 10, 315-323.

Critchley C. N. R. and Poulton S. M. C. (1998) A method to optimize precision and scale in grassland monitoring. Journal of Vegetation Science 9, 837-846.

Csurhes S. M. and Edwards R. (1998) Potential Environmental Weeds in Australia: Candidate Species for Preventative Control. Biodiversity Group, Environment Australia, Canberra.

Cunningham G. M., Mulham W. E., Milthorpe P. L. and Leigh J. (1981) Plants of Western New South Wales. NSW Government Printing Office, Sydney.

D'Antonio C. M. (2000) Fire, plant invasions, and global change. In: Invasive Species in a Changing World (eds H. A. Mooney and R. J. Hobbs) pp. 65-93. Island Press, Washington.

D'Antonio C. M., Hughes R. F., Mack M., Hitchcock D. and Vitousek P. M. (1998) The response of native species to removal of invasive exotic grasses in a seasonally dry Hawaiian woodland. Journal of Vegetation Science 9, 699-712.

D'Antonio C. M., Hughes R. F. and Vitousek P. M. (2001) Factors influencing dynamics of two invasive C4 grasses in seasonally dry Hawaiian woodlands. Ecology 82, 89-104.

D'Antonio C. M., Tunison J. T. and Loh R. K. (2000) Variation in the impact of exotic grasses on native plant composition in relation to fire across an elevation gradient in Hawaii. Austral Ecology 25, 507-522.

D'Antonio C. M. and Vitousek P. M. (1992) Biological invasions by exotic grasses, the grass/fire cycle, and global change. Annual Review of Ecology and Systematics 23, 63-87.

Daehler C. C. (1998) The taxonomic distribution of invasive angiosperm plants: ecological insights and comparison to agricultural weeds. Biological Conservation 84, 167-180.

Daehler C. C. and Carino D. A. (1998) Recent replacement of native pili grass (Heteropogon contortus) by invasive African grasses in the Hawaiian Islands. Pacific Science 52, 220-227.

Daehler C. C. and Strong D. R. (1994) Native plant biodiversity vs the introduced invaders: status of the conflict and future management options. In: Biological Diversity: Problems and Challenges (eds S. K. Majumbar, E. J. Brenner, J. E. Lovich, J. F. Schalles and E. W. Miller) pp. 92-113. Pennsylvania Academy of Science, Easton, Pennsylvania.

Daehler C. C. and Strong D. R. (1996) Status, prediction and prevention of introduced cordgrass Spartina spp. invasions in Pacific estuaries, USA. Biological Conservation 78, 51-58.

Daubenmire R. (1968) Ecology of fire in grasslands. Advances in Ecological Research 5, 209-266.

Davidson R. L. (1951) cited in Roux E. (1969), Grass. A Story of Frankenwald. Oxford

191

University Press, Cape Town, South Africa.

DiTomaso J. M., Kyser G. B. and Hastings M. S. ( 1999) Prescribed burning for the control of yellow starthistle (Centaurea solstitialis) and enhanced native plant diversity. Weed Science 47, 233-242.

DNR (1998) The Desert Uplands. The State of Queensland. Land Planning and Management, Brisbane.

Douglas M. M., Townsend S. and Lake P. S. (1996) The role of fire in the management of tropical savanna streams. In: The Future of Tropical Savannas: an Australian Perspective, Poster Papers (ed. J. Brown) pp. 13-15. CSIRO, Australia.

Downey P. O. and Smith J. M. B. (2000) Demography of the invasive shrub Scotch Broom (Cytisus scoparius) at Barrington Tops, New South Wales: insights for management. Austral Ecology 25, 477-485.

Dunbar K. R. and Facelli J. M. (1999) The impact of a novel invasive species, Orbea variegata (African carrion flower), on the chenopod shrublands of South Australia. Journal of Arid Environments 41, 37-48.

Dyer R. (2000) The role of fire on pastoral lands in northern Australia. In: Fire and Sustainable Agricultural and Forestry Development in Eastern Indonesia and Northern Australia. Proceedings of an International Workshop Held at Northern Territory University, Darwin, Australia, 13-15 April 1999, (eds J. Russell-Smith, G. Hill, S. Djoeroemana, and B. Myres) The Australian Centre for International Agricultural Research, Canberra.

Dyer R., Craig A. and Grice A. C. (1997) Fire in northern pastoral lands. In: Fire in the Management of Northern Australian Pastoral Lands pp. 24-40. The Tropical Grassland Society of Australia, Brisbane, Australia.

Eller K. G. (1994) cited in Alexander J. M. and D'Antonio C. M. (2003), Seed bank dynamics of french broom in coastal California Grasslands: effects of stand age and prescibed burning on control and restoration. Restoration Ecology 11, 185-197.

Elton C. S. (1958) cited in Barrow (1995), The Ecology and Management of Gamba Grass (Andropogon gayanus Kunth.). Northern Territory Department of Primary Industry and Fisheries, Northern Territory.

Engle D. M., Palmer M. W., Crockett J. S., Mitchell R. L. and Stevens R. (2000) Influence of late season fire on early successional vegetation of an Oklahoma prairie. Journal of Vegetation Science 11, 135-144.

Enright N. J., Goldblum D., Ata P. and Ashton D. H. (1997) The independent effects of heat, smoke and ash on emergence of seedlings from the soil seed bank of a heathy Eucalyptus woodland in Grampians (Gariwerd) National Park, western Victoria. Australian Journal of Ecology 22, 81-88.

Enright N. J. and Lamont B. B. (1989) Seed banks, fire season, safe sites and seedling recruitment in five co-occurring Banksia species. Journal of Ecology 77, 1111-1122.

Ernst W. H. O. (1991) Fire, dry heat and germination of savanna grasses in Botswana. In: Modern Ecology: Basic and Applied Aspects (eds G. Esser and D. Overdieck) pp. 349-361. Elsevier, Amsterdam.

Eser U. (1998) Assessment of plant invasions: theoretical and philosophical fundamentals.

192

In: Plant Invasions: Ecological Mechanisms and Human Responses (eds U. Starfinger, K. Edwards, I. Kowarik and M. Williamson) pp. 95-107. Backhuys Publishers, Leiden, The Netherlands.

Evans R. A. and Young J. A. (1970) Plant litter and establishment of alian annual weed species in rangeland communities. Weed Science 18, 697-703.

Evans R. D., Rimer R., Sperry L. and Belnap J. (2001) Exotic plant invasion alters nitrogen dynamics in an arid grassland. Ecological Applications 11, 1301-1310.

Everson T.M. (1999) Seed dynamics of Themeda triandra in montane grasslands of South Africa. In: People and Rangelands: Building the Future. Proceedings of the VI International Rangeland Congress, Townsville, Queensland, 19-23 July 1999, (eds D. Eldridge and D. Freudenberger) pp. 301-302. International Rangeland Congress, Inc., Aitkenvale, Australia.

Facelli J. M. (1994) Multiple indirect effects of plant litter affect the establishment of woody seedlings in old fields. Ecology 75, 1727-1735.

Facelli J. M. and Kerrigan R. (1996) Effects of ash and four types of litter on the establisment of Eucalyptus obliqua. Ecoscience 3, 319-324.

Facelli J. M. and Pickett S. T. A. (1991a) Plant litter: its dynamics and effects on plant community structure. The Botanical Review 57, 1-32.

Facelli J. M. and Pickett S. T. A. (1991b) Plant litter: light interception and effects on an old-field community. Ecology 72, 1024-1031.

Facelli J. M., Williams R., Fricker S. and Ladd B. (1999) Establishment and growth of seedlings of Eucalypyus obliqua: interactive effects of litter, water and pathogens. Australian Journal of Ecology 24, 484-494.

Fairfax R. J. and Fensham R. J. (2000) The effects of exotic pasture development on floristic diversity in central Queensland, Australia. Biological Conservation 94, 11-21.

Fensham R. J. (1996) Land clearance and conservation of inland dry rainforest in north Queensland, Australia. Biological Conservation 75, 289-298.

Ferrandis P., Herranz J. M. and Martinez-Sanchez J. J. (1999) Fire impact on a maquis soil seed bank in Cabaneros National Park (central Spain). Israel Journal of Plant Sciences 47, 17-26.

Fowler N. L. (1986) Microsite requirements for germination and establishment of three grass species. The American Naturalist 115, 131-145.

Fowler N. L. (1988) What is a safe site?: neighbour, litter, germination date, and patch effects. Ecology 69, 947-961.

Fox M. D. (1995) Conserving biodiversity: impact and management of exotic organisms. In: Conserving Biodiversity: Threats and Solutions (eds R. A. Bradstock, T. D. Auld, D. A. Keith, R. T. Kingsford, D. Lunney and D. P. Sivertsen) pp. 177-183. Surrey Beatty and Sons, Chipping Norton, NSW, Australia.

Franklin T., Asher J. and Barclay E. (1999) Invasion of the aliens: exotic plants impact wildlife. Wildlife Society Bulletin 27, 873-875.

Franks A. J. (2002) The ecological consequences of buffel grass Cenchrus ciliaris

193

establishment within remnant vegetation of Queensland. Pacific Conservation Biology 8, 99-107.

Franks A. J., Butler D and Fairfax R. (2000) A weed by any other name. Wildlife Australia 37, 24.

Franks A. and Hannah, D. (2001) When is a weed not a weed? The ecological consequences of buffel grass establishment in remnant vegetation and native pastures. In: Remnant Vegetation in the Brigalow Belt - Management and Conservation. A Symposium Presented by Centre for Environmental Management, Central Queensland University, Session 4. Central Queensland University, Rockhampton, Australia.

Freifelder R. R., Vitousek P. M. and D'Antonio C. M. (1998) Microclimate change and effect on fire following forest-grass conversion in seasonally dry tropical woodland. Biotropica 30, 286-297.

French K. and Eardley K. (1997) The impact of weed infectations on litter invertebrates in coastal vegetation. In: Frontiers in Ecology. Building the Links (eds N. Klomp and I. Lunt) pp. 89-102. Elsevier Science, Oxford.

Friday K. S., Drilling M. E. and Garrity D. P. (1999) cited in Chikoye, D and Ekeleme, F (2001), Weed flora and soil seed banks in fields dominated by Imperata cylindrica in the moist savannah of West Africa. Weed Research 41, 475-490.

Furbush P. B. (1953) cited in Daubenmire R E (1968), Ecology of fire in grasslands. Advances in Ecological Research 5, 209-266.

Gardener C. J. (1980) Tolerance of perennating Stylosanthes plants to fire. Australian Journal of Agriculture and Animal Husbandry 20, 587-593.

Gardener M. R. and Sindel B. M. (1998) The biology of Nassella and Achnatherum species naturalized in Australia and the implications for management on conservation lands. Plant Protection Quarterly 13, 76-79.

Genstat (2001) Release 4.21, 5th edition. Lawes Agricultural Trust (Rothamsted Experimental Station), Britain.

Gill A. M. (1975) Fire and the Australian flora. Australian Forestry 38, 4-25.

Gill A. M. (1981) Adaptive responses of Australian vascular plant species to fires. In: Fire and the Australian Biota (eds A. M. Gill, R. H. Grove and I. R. Noble) pp. 243-271. Australian Academy of Science, Canberra.

Gill A. M., Hoare J. R. L. and Cheney N. P. (1990) Fires and their effects in the wet-dry tropics of Australia. In: Fire in the Tropical Biota: Ecological Processes and Global Change (ed. J. G. Goldammer) pp. 159-178. Springer Verlag, Berlin.

Gill A.M. and Knight, I. K. (1991) Fire measurement. In: Conference on Bushfire Modelling and Fire Danger Rating Systems. Proceedings, pp. 137-146. CSIRO, Canberra, ACT.

Gill A. M., Moore P. H. R. and Williams R. J. (1996) Fire weather in the wet-dry tropics of the world heritage Kakadu National Park, Australia. Australian Journal of Ecology 21, 302-308.

Gleadow R. M. and Ashton D. H. (1981) Invasion by Pittosporum undulatum of the forests of central Victoria. I. Invasion patterns and plant morphology. Australian Journal of Botany

194

29, 705-720.

Glenn-Lewin D. C., Johnson L. A., Jurik T. W., Akey A., Leoschke M. and Rosburg T. (1990) Fire in central North American grasslands: vegetative reproduction, seed germination, and seedling establishment. In: Fire in North American Tallgrass Prairies (eds S. L. Collins and L. L. Wallace) pp. 28-45. University of Oklahoma Press, Norman.

Gordon D. R. (1998) Effects of invasive, non-indigenous plant species on ecosystem processes: lessions from Florida. Ecological Applications 8, 975-989.

Grace J. B. (1999) The factors controlling species density in herbaceous plant communities: an assessment. Perspectives in Plant Ecology, Evolution and Systematics 2, 1-28.

Grice A. C. (1997) Post-fire regrowth and survival of the invasive tropical shrubs Cryptostegia grandiflora and Ziziphus mauritiana. Australian Journal of Ecology 22, 49-55.

Grice A. C. (2000) Weed management in Australian rangelands. In: Australian Weed Management Systems (ed. B. M. Sindel) pp. 431-458. R.G. and F.J. Richardson, Meredith, Victoria.

Grice A. C., Campbell, S. D. and Vitelli, J. S. (2000) The principles and practice of weed control for northern Australia. In: Northern Grassy Landscapes Conference, Katherine, Northern Territory 29-31 August 2000, Cooperative Research for the Sustainable Development of Tropical Savannas, Darwin, Northern Territory.

Grice A. C. and Slatter S. M. (1997) Fire in the Management of Northern Australian Pastoral Lands. Proceedings of a Workshop Held in Townsville, Queensland, Australia, 1996. The Tropical Grassland Society of Australia Inc., Brisbane, Australia.

Griffin G. F. (1993) The spread of buffel grass in inland Australia: land use conflicts. In: Proceedings of the 10th Australian Weeds Conference and the 14th Asian Pacific Weed Science Society Conference, September 1993, Brisbane, Australia, pp. 501-504. Weed Society of Queensland, Brisbane, Australia.

Griffin G. F. and Friedel M. H. (1984a) Effects of fire on central Australian rangelands. II Changes in tree and shrub populations. Australian Journal of Ecology 9, 395-403.

Griffin G. F. and Friedel M. H. (1984b) Effects of fire on central Australian rangelands. I Fire and fuel characteristics and changes in herbage and nutrients. Australian Journal of Ecology 9, 381-393.

Griffin G. F. and Hodgkinson K. C. (1986) cited in Dyer R., Craig A. and Grice A. (1997), Fire in northern pastoral lands. In: Fire in the Management of Northern Australian Pastoral Lands pp. 24-40. The Tropical Grassland Society of Australia, Brisbane, Australia.

Griffin G. F., Stafford Smith D. M., Morton S. R., Allan G. E. and Masters K. A. (1989) Status and implications of the invasion of Tamarisk (Tamarix aphylla) on the Finke River, Northern Territory, Australia. Journal of Environmental Management 29, 297-315.

Grigg A., Shelton M. and Mullen B. (2000) The nature and management of rehabilitated pastures on open-cut coal mines in central Queensland. In: Pastures for Production and Protection. Tropical Grassland Conference, Emerald 26-28 April 2000 pp. 70-76. The Tropical Grassland Society of Australia, St Lucia, Queensland.

Grime J. P. (1973) Control of species density in herbaceous vegetation. Journal of Environmental Management 1, 151-167.

195

Grime J. P. (1979) Plant Strategies and Vegetation Processes. John Wiley and Sons, Chichester.

Grime J. P. (1989) Seed banks in ecological perspective. In: Ecology of Soil Seed Banks (eds M. A. Leck, V. T. Parker and R. L. Simpson) pp. xv-xxii. Academic Press, Inc., San Diego, California.

Gross K. L. (1990) A comparison of methods for estimating seed numbers in the soil. Journal of Ecology 78, 1079-1093.

Gross K. L. and Werner P. A. (1982) Colonizing abilities of "biennial" plant species in relation to ground cover: implications for their distributions in a successional sere. Ecology 64, 921-931.

Groves R. H. (1986a) Invasion of mediterranean ecosystems by weeds. In: Resilience in Mediterranean-Style Ecosystems (eds A. J. M. Hopkins and B. B. Lamont) pp. 129-145. Dr W. Junk Publishers, Dordrecht.

Groves R. H. (1986b) Plant invasions of Australia: an overview. In: Ecology of Biological Invasions: An Australian Perspective (eds R. H. Groves and J. J. Burden) pp. 137-149. Australian Academy of Science, Canberra.

Groves R. H. (1991) Control of environmental weeds. Plant Protection Quarterly 6, p95-98.

Groves R. H. and Willis A. J. (1999) Environmental weeds and the loss of native plant biodiversity: some Australian examples. Australian Journal of Environmental Management 6, 164-171.

Grubb P. J. (1977) The maintenance of species-richness inplant communities: the importance of the regeneration niche. Biological Review 52, 107-145.

Hacker J. B. (1989) The potential for buffel grass renewal from seed in 16-year-old buffel grass-siratro pastures in south-east Queensland. Journal of Applied Ecology 26, p213-222.

Hacker J. B. and Ratcliff A. (1989) Seed dormancy and factors controlling dormancy breakdown in buffel grass accessions from contrasting provenances. Journal of Applied Ecology 26, 201-212.

Hall T. J. (2001) History and development of buffel grass pasture lands in Queensland. In: Buffel Grass Symposium. Proceedings of Workshop Held at Theodore on 21-23 February 2000 (eds J. Kyte, H. Lees, I. Loxton and B. Cook) pp. 2-12. Department of Primary Industries, Brisbane, Australia.

Hamilton W. T. and Scifres C. J. (1982) Prescribed burning during winter for maintenance of buffel grass. Journal of Range Management 35, 9-12.

Hamrick J. L. and Lee J. M. (1987) Effect of soil surface topography and litter cover on the germination, survival, and growth of musk thistle (Caduus nutans). American Journal of Botany 74, 451-457.

Harden G. J. (1993) Flora of New South Wales. Volume 4. NSW University Press, Kensington, N.S.W.

Harper J. L. (1977) Population Biology of Plants. Academic Press, England.

Harper J. L., Williams J. T. and Sagar G. R. (1965) The behaviour of seeds in soil. 1. The

196

heterogeneity of soil surface and its role in determining the establishment of plants from seed. Journal of Ecology 53, 237-286.

Harrington G. N. and Ross I. C. (1974) The savanna ecology of Kipepo Valley National Park. East African Wildlife Journal 12, 93-105.

Harwood M. R., Hacker J. B. and Mott J. J. (1999) Field evaluation of seven grasses for use in the revegetation of lands disturbed by coal mining in central Queensland. Australian Journal of Experimental Agriculture 39, 307-16.

Heady H. F. (1977) cited in Vlok J. H. J. (1988), Alpha diversity of lowland fynbos herbs at various levels of infestation by alien annuals. African Journal of Botany 54, 623-627.

Hedge P. and Kriwoken L. K. (2000) Evidence for effects of Spartina anglica invasion on benthic macrofauna in Little Swanport estuary, Tasmania. Austral Ecology 25, 150-159.

Heyligers P. C. (1985) The impact of introduced plants on foredune formation in south-eastern Australia. Proceedings of the Ecological Society of Australia 14, 23-41.

Heywood V. H. (1989) Patterns, extents and modes of invasions by terrestrial plants. In: Biological Invasions: A Global Perspective (eds J. A. Drake and H. A. Mooney) pp. 31-60. Published on behalf of the Scientific Committee on Problems of the Environment of the International Council of Scientific Unions by Wiley, Chichester.

Hitchmough J. D., Kilgour R. A., Morgan J. W. and Shears I. G. (1994) Efficacy of some grass specific herbicides in controlling exotic grass seedlings in native grassy vegetation. Plant Protection Quarterly 9, 28-34.

Hoare J. (1985) A model of fire behaviour and resulting damage to vegetation for predicting the biological effects of fire on plant communities at Kakadu National Park. Tech. Memo. CSIRO Div. Land and Water Resource 85, 138-152.

Hobbs R. J. and Atkins L. (1988) Effects of disturbance and nutrient addition on native and introducted annuals in plant communities in the Western Australian wheatbelt. Australian Journal of Ecology 13, 171-179.

Hodgkinson K. C., Harrington G. N., Griffin G. F., Noble J. C. and Young M. D. (1984) Management of vegetation with fire. In: Management of Australia's Rangelands (eds G. N. Harrington, A. D. Wilson and M. D. Young) pp. 141-156. CSIRO, Melbourne, Australia.

Holl K. D., Steele H. N., Fusari M. H. and Fox L. R. (2000) Seed banks of maritime chaparral and abandoned roads: potential for vegetation recovery. Journal of the Torrey Botanical Society 127, 207-220.

Holm L. G., Plucknett D. L., Pancho J. V. and Herberger J. P. (1977) The World's Worst Weeds. University Press of Hawaii for the East-West Center, Honolulu.

Holmes P. M. and Cowling R. M. (1997) Diversity, composition and guild structure relationships between soil-stored seed banks and mature vegetation in alien plant-invaded South African fynbos shrublands. Plant Ecology 133, 107-122.

Houston W. A. and Duivenvoorden L. J. (2002) Replacement of littoral native vegetation with the ponded pasture grass Hymenachne amplexicaulis: effects on plant, macroinvertebrate and fish biodiversity of backwaters in the Fitzroy River, Central Queensland, Australia. Marine and Freshwater Research 53, 1235-1244.

197

Howden S. M. (1988) Some Aspects of the Ecology of Four Tropical Grasses With Species Emphasis on Bothriochloa pertusa. Thesis, Division of Australian Environmental Studies, Griffith University.

Howe H. F. (1994) Response of early- and late-flowering plants to fire season in experimental prairies. Ecological Applications 4, 121-133.

Hughes F., Vitousek P. M. and Tunison T. (1991) Alien grass invasion and fire in the seasonal submontane zone of Hawaii. Ecology 72, 743-746.

Hulbert L. C. (1988) Causes of fire effects in tallgrass prairie. Ecology 69, 46-58.

Humphreys L. R. (1967) Buffel grass (Cenchrus ciliaris) in Australia. Tropical Grasslands 1, 123-134.

Humphreys L. R. (1974) A Guide to Better Pastures for the Tropics and Sub-Tropics. 3rd edition. Wright Stephenson and Co, Melbourne.

Humphreys L. R. (1981) Tropical Seed Production. 2nd edition. Food and Agriculture Organisation of the United Nations, Rome.

Humphreys L. R. and Partridge I. (1995) A Guide to Better Pastures for the Tropics and Sub-Tropics. Revised 5th edition. NSW Agriculture, Newcastle, N.S.W.

Humphries F. R. and Craig F. G. (1981) Effects of fire on soil chemical, structural and hydrological properties. In: Fire and the Australian Biota (eds A. M. Gill, R. H. Grove and I. R. Noble) pp. 177-202. Australian Academy of Science, Canberra.

Humphries S. E., Mitchell D. S. and Groves R. H. (1991) Plant Invasions of Australian Ecosystems: A Status Review and Management Directions. CSIRO, Canberra.

Hyams D. (2003) CurveExpert 1.3. A Comprehensive Curve Fitting System for Windows.

Isbell R. F. (1996) The Australian Soil Classification. CSIRO Publishing, Melbourne.

Jacobs W. L. (1981) Man's impact on native vegetation. Australian Natural History 20, 207-210.

Johnson R. W. (1986) Weeds in the functioning of tropical plant communities. In: Tropical Plant Communities (eds H. T. Clifford and R. L. Specht) pp. 104-112. Department of Botany, University of Queensland, St Lucia, Brisbane.

Keeley J. E., Morton B. A., Pedrosa A. and Trotter P. (1985) Role of allelopathy, heat and charred wood in the germination of chaparral herbs and suffrutescents. Journal of Ecology 73, 445-458.

Keighery G. J. (1991) Environmental weeds of Western Australia. In: Plant Invasions: The Incidence of Environmental Weeds in Australia (ed. Australian National Parks and Wildlife Service) pp. 180-188. National Parks & Wildlife Service, Canberra.

Kellman M., Miyanishi K. and Hiebert P. (1985) Nutrient retention by savanna ecosystems. II. Retention after fire. Journal of Ecology 73, 953-962.

Kemp P. R. (1989) Seed banks and vegetation processes in deserts. In: Ecology of Soil Seed Banks (eds M. A. Leck, V. T. Parker and R. L. Simpson) pp. 257-281. Academic Press, Inc., San Diego, California.

198

Kendle A. D. and Rose J. E. (2000) The aliens have landed! What are the justifications for ‘native only’ policies in landscape plantings? Landscape and Urban Planning 47, 19-31.

Kennard D. G. and Walker B. H. (1973) Relationships between tree canopy cover and Panicum maximum in the vicinity of Fort Victoria. Rhodesian Journal of Agricultural Research 11, 145-153.

Kilburn P. D. (1966) Analysis of the species-area relation. Ecology 47, 831-843.

Kloot P. M. (1983) The role of common iceplant (Mesembryanthemum crystallinum) in the deterioration of medic pastures. Australian Journal of Ecology 8, 301-306.

Knapp A. K. and Seastedt T. R. (1986) Detritus accumulation limits productivity of tallgrass prairie. Bioscience 36, 662-668.

Lacey J. R., Marlow C. B. and Lane J. R. (1989) Influence of spotted knapweed (Centaurea maculosa) on surface runoff and sediment yield. Weed Technology 3, 627-631.

Lambrinos J. G. (2000) The impact of the invasive alien grass Cortaderia jubata (Lemoine) Stapf on an endangered Mediterranean-type shrubland in California. Diversity and Distributions 6, 217-231.

Lane A. M. and Williams, R. J. (1997) The effects of wet season burning on understorey vegetation in a humid tropical savanna in northern Australia. In: Bushfire '97. Proceedings. Australian Bushfire Conference, 1997, Darwin, (eds B. J. McKaige, R. J. Williams, and W. M. Waggitt) CSIRO Tropical Ecosystems Research Centre.

Lane P. W. and Payne R. W. (1998) Genstat for Windows: an Introductory Course. Lawes Agricultural Trust (Rothamsted Experimental Station), Harpenden, UK.

Laterra P. (1997) Post-burn recovery in the flooding Pampa: impact of an invasive legume. Journal of Range Management 50, p274-277.

Latz P. K. (1991) Buffel and couch grass in Central Australian creeks and rivers. Newsletter of the Central Australian Conservation Council. Newsletter of the Central Australian Conservation Council 5.

Lauenroth W. K. and Aguilera M. O. (1998) Plant-plant interactions in grasses and grasslands. In: Population Biology of Grasses (ed. G. P. Cheplick) pp. 209-230. Cambridge University Press, Cambridge.

Lauenroth W. K., Sala O. E., Coffin D. P. and Kirchner T. B. (1994) The importance of soil water in the recruitment of Bouteloua gracilis in the shortgrass steppe. Ecological Applications 4, 741-749.

Lawrence E. (1995) Henderson's Dictionary of Biological Terms. Longman Group UK Limited, Essex, England.

Lazarides M., Cowley K. J., Hohnen P. and CSIRO (1997) CSIRO Handbook of Australian Weeds. CSIRO Publishing, East Melbourne.

Leigh J. H. and Noble J. C. (1981) The role of fire in the management of rangelands in Australia. In: Fire and the Australian Biota (eds A. M. Gill, R. H. Grove and I. R. Noble) pp. 471-495. Australian Academy of Science, Canberra.

Leigh J. H. and Briggs J.D. (1992) cited in Adair (1995), The threat of environmental weeds

199

to biodiversity in Australia: a search for solutions. In: Conserving Biodiversity: Threats and Solutions (eds R. A. Bradstock, T. D. Auld, D. A. Keith, R. T. Kingsford, D. Lunney and D. P. Sivertsen) pp. 184-201. Surrey Beatty and Sons, Chipping Norton, NSW, Australia.

Leigh J., Boden R. and Briggs J. D. (1984) Extinct and Endangered Plants of Australia. Macmillan, South Melbourne, Victoria.

Lesica P. and DeLuca T. H. (1996) Long-term harmful effects of crested wheatgrass on Great Plains grassland ecosystems. Journal of Soil and Water Conservation 51, 408-409.

Levine J. M., Vilà M., D'Antonio C. M., Dukes J. S., Grigulis K. and Lavoral S. (2003) Mechanisms underlying the impacts of exotic plant invasions. The Royal Society of London Series B-Biological Sciences 270, 775-781.

Ley R. E. and D'Antonio C. M. (1998) Exotic grass invasion alters potential rates of N fixation in Hawaiian woodlands. Oecologia 113, 179-187.

Lippincott C. L. (2000) Effects of Imperata cylindrica (L.) Beauv. (Cogongrass) invasion on fire regime in Florida sandhill (USA). Natural Areas Journal 20, 140-149.

Loch A. (1993) Tropical pasture establishment. 5. Improved handling of chaffy grass seeds: options, opportunities and value. Tropical Grasslands 27, 314-326.

Lonsdale W.M. (1992) The impact of weeds in national parks. In: International Weed Control Congress: Proceedings of the First International Weed Control Congress, 17-21 February, 1992, Monash University, Australia, (ed. J. H. Combellack) pp. 145-149. Weed Science Society of Victoria, Melbourne.

Lonsdale W. M. (1994) Inviting trouble: introduced pasture species in northern Australia. Australian Journal of Ecology 19, 345-354.

Lonsdale W. M. and Braithwaite R. W. (1991) Assessing the effects of fire on vegetation in tropical savannas. Australian Journal of Ecology 16 , 363-374.

Loope L. L., Sanchez P. G., Tarr P. W., Loope W. L. and Anderson R. L. (1988) Biological invasions of arid land nature reserves. Biological Conservation 44, 95-118.

López-Barrera F. and González-Espinosa M. (2001) Influence of litter on emergence and early growth of Quercus rugosa: a laboratory study. New Forests 21, 59-70.

Low T. (1997) Tropical pasture plants as weeds. Tropical Grasslands 31, 337-343.

Loyn R. H. and French K. (1991) Birds and environmental weeds in south-eastern Australia. Plant Protection Quarterly 6, 137-149.

Ludwig J. A., Eager R. W., Liedloff A. C., McCosker J. C., Hannah D., Thurgate N. Y., Woinarski J. C. Z. and Catterall C. P. (2000) Clearing and grazing impacts on vegetation patch structures and fauna counts in eucalypt woodland, Central Queensland. Pacific Conservation Biology 6, 254-272.

Lunt I. D. (1990) Impact of an autumn fire on a long-grazed Themeda triandra (kangaroo grass) grassland: implications for the management of invaded, remnant vegetations. Victorian Naturalist 107, 45-51.

Lunt I. D. (1994) Variation in flower production of nine grassland species with time since fire, and implications for grassland management and restoration. Pacific Conservation

200

Biology 1, 359-366.

Lunt I. D. and Morgan J. W. (2002) The role of fire regimes in temperate lowland grasslands of southeastern Australia. In: Flammable Australia. The Fire Regimes and Biodiveristy of a Continent (eds R. A. Bradstock, R. A. Williams and M. A. Gill) pp. 178-196. Cambridge University Press, Cambridge, New York.

Macdonald I. A. W. and Frame G. W. (1988) The invasion of introduced species into nature reserves in tropical savannas and dry woodlands. Biological Conservation 44, 67-93.

Macdonald I. A. W., Loope L. L., Usher M. B. and Hamann O. (1989) Wildlife conservation and the invasion of nature reserves by introduced species: a global perspective. In: Biological Invasions: A Global Perspective (eds J. A. Drake and H. A. Mooney) pp. 215-255. Published on behalf of the Scientific Committee on Problems of the Environment of the International Council of Scientific Unions by Wiley, Chichester, New York.

Macdonald I. A. W. (1983) cited in Daehler C. C. and Strong D. R. (1994), Native plant biodiversity vs the introduced invaders: status of the conflict and future management options. In: Biological Diversity: Problems and Challenges (eds S. K. Majumbar, E. J. Brenner, J. E. Lovich, J. F. Schalles and E. W. Miller) pp. 92-113. Pennsylvania Academy of Science, Easton, Pennsylvania.

Mack M. C., D'Antonio C. M. and Ley R. E. (2001) Alteration of ecosystem nitrogen dynamics by exotic plants: a case study of C4 grasses in Hawaii. Ecological Applications 11, 1323-1335.

Mack R. N. (1996) Predicting the identity and fate of plant invaders: emergent and emerging approaches. Biological Conservation 78, 107-121.

Mack R. N. and D'Antonio C. M. (1998) Impacts of biological invasions on disturbance regimes. Trends in Ecology and Evolution 13, 195-198.

MacNally R. and Quinn G. P. (1998) Symposium introduction: the importance of scale in ecology. Australian Journal of Ecology 23, 1-7.

Main M. B. and Barry M. J. (2002) Influence of season of fire on flowering of wet prairie grasses in South Florida, USA. Wetlands 22, 430-434.

Manchester S. J. and Bullock J. M. (2000) The impacts of non-native species on UK biodiversity and the effectiveness of control. Journal of Applied Ecology 37, 845-864.

Mannetje L. ’t, Cook S. J. and Wildin J. H. (1983) The effects of fire on a buffel grass and Siratro pasture. Tropical Grasslands 17, 30-39.

Marriott S. (1955) Buffel grass. Journal of the Australian Institute of Agricultural Science 21, 277.

Masters R. A. and Sheley R. L. (2001) Principles and practices for managing rangeland invasive plants. Journal of Range Management 54, 502-517.

McArthur (1972) Fire control in the arid and semi-arid lands of Australia. In: The Use of Trees and Shrubs in the Dry Country of Australia (ed. N. Hall) pp. 488-515. Australian Gov. Printing Service, Canberra.

McArthur A. G. and Cheney N. P. (1966) The characterization of fires in relation to ecological studies. Australian Forest Research 2, 36-45.

201

McCallum B. (1998) An Investigation of Native Tree Incursion into Native Grassland at Moorrinya National Park. Honours Thesis, James Cook University, Townsville.

McIntyre S., Ladiges P. Y. and Adams G. (1988) Plant species richness and invasion by exotics in relation to disturbance of wetland communities on the Riverine Plain, NSW. Australian Journal of Ecology 13, 361-373.

McIntyre S. and Lavorel S. (1994) Predicting richness of native, rare and exotic plants in response to habitat and disturbance variables across a variegated landscape. Conservation Biology 2, 521-531.

McIvor J. G. (1984) Phosphorus requirements and responses of tropical pasture species: native and introduced grasses, and introduced legumes. Australian Journal of Experimental Agriculture and Animal Husbandry 24, 370-378.

McIvor J. G. (1987) Changes in germinable seed levels in soil beneath pastures near Townsville, North Queensland. Australian Journal of Experimental Agriculture 27, 283-289.

McIvor J. G. (1998) Pasture management in semi-arid tropical woodlands: effects on species diversity. Australian Journal of Ecology 23, 349-364.

McIvor J. G. (2003) Competition affects survival and growth of buffel grass seedlings - is buffel grass a coloniser or an invader? Tropical Grasslands 37, 176-181.

McIvor J. G., Ash A. J. and Cook G. D. (1995) Land condition in the tropical tallgrass pasture lands. 1. Effects on herbage production. Rangeland Journal 17, p69-85.

McIvor J. G., Ash, A. J. and Grice, A. C. (2000) Introduced grasses: do they add value or should they be vilified? In: Northern Grassy Landscapes Conference, Katherine, Northern Territory 29-31 August 2000., Cooperative Research for the Sustainable Development of Tropical Savannas, Darwin, Northern Territory.

McIvor J. G. and Gardener C. J. (1981) Establishment of introduced grasses at different stages of pasture development: effects of seedbed. Australian Journal of Experimental Agriculture and Animal Husbandry 21, 417-423.

McIvor J. G. and Gardener C. J. (1985) Germination of introduced pasture species in the semi-arid tropics. In: Ecology and Management of the World's Savannas (eds J. C. Tothill and J. J. Mott) pp. 277-279. Australian Academy of Science, Canberra, Australia.

McIvor J. G. and Gardener C. J. (1991) Soil seed densities and emergence patterns in pastures in the seasonally dry tropics of northeastern Australia. Australian Journal of Ecology 16, 159-169.

McIvor J. G. and Gardener C. J. (1994) Germinable soil seed banks in native pastures in north-eastern Australia. Australian Journal of Experimental Agriculture 34, 1113-1119.

McIvor J. G., Gardener C. J. and Spain A. V. (1991) A Description of the ECOSSAT Experimental Site at Hillgrove Near Charters Towers, North Queensland. Tropical Agronomy Technical Memorandum No. 70. CSIRO Division of Tropical Crops and Pastures, Brisbane.

McIvor J. G., Jones R. M. and Taylor J. (1993) Tropical pasture establishment. 4. Population dynamics of sown species in developing pastures. Tropical Grasslands 27, 302-313.

McIvor J. G. and McIntyre S. (1997) Responsible use of exotic tropical pasture cultivars -

202

an ecological framework. Tropical Grasslands 31, 332-336.

McIvor J. G. and Orr D. M. (1991) Sustaining productive pastures in the tropics. 3. Managing native grasslands. Tropical Grasslands 25, 91-97.

McMahon A. R. G., Carr, G. W., Bedggood, S. E., Hill, R. J. and Pritchard, A. M. (1996) 8. Prescribed fire and control of coast wattle (Acacia sophorae (Labill.) R.Br.) invasion in coastal heath south-west Victoria. In: Fire and Biodiversity. The Effects and Effectiveness of Fire Management. Proceedings of the Conference Held 8-9 October, Footscray, Melbourne., Department of the Environment, Sport and Territories, Biodiversity Unit, Canberra, ACT.

Michael P. W. (1994) Alien plants. In: Australian Vegetation (ed. R. H. Groves) pp. 57-83. Cambridge University Press, Cambridge.

Mittelbach G. G., Steiner C. F., Scheiner S. M., Gross K. L., Reynolds H. L., Waide R. B., Willig M. R., Dodson S. I. and Gough L. (2001) What is the observed relationship between species richness and productivity? Ecology 82, 2381-2396.

Mooney H. A., Hamburg S. P. and Drake J. A. (1986) cited in Woods K. D. (1997), Community response to plant invasion. In: Assessment and Management of Plant Invasions (eds J. O. Luken and J. W. Thieret) pp. 56-68. Springer-Verlag, New York.

Moore D. R. J. and Keddy P. A. (1989) The relationship between species richness and standing crop in wetlands: the importance of scale. Vegetatio 9, 99-106.

Morgan J. W. (1998) Composition and seasonal flux of the soil seed bank of species-rich Themeda triandra grasslands in relation to burning history. Journal of Vegetation Science 9, 145-156.

Morgan J. W. (1999) Defining grassland fire events and the response of perennial plants to annual fire in temperate grasslands of south-eastern Australia. Plant Ecology 144, 127-144.

Morrison M. L. (1997) Experimental design for plant removal and restoration. In: Assessment and Management of Plant Invasions (eds J. O. Luken and J. W. Thieret) pp. 104-116. Springer-Verlag, New York.

Mott J. J. (1982) Fire and survival of Stylosanthes spp. in the dry savanna woodlands of the Northern Territory. Australian Journal of Agricultural Research 33, 203-211.

Mott J. J. (1986) Planned invasions of Australian tropical savannas. In: Ecology of Biological Invasions: an Australian Perspective (eds R. H. Groves and J. J. Burden) pp. 89-96. Australian Academy of Science, Canberra.

Mott J. J. and Andrew M. H. (1985a) Effect of fire on the grass understorey of Australian savannas. Tech. Memo. CSIRO Div. Land and Water Resource 85, 77-91.

Mott J. J. and Andrew M. H. (1985b) The effect of fire on the population dynamics of native grasses in tropical savannas of north-west Australia. Proceedings of the Ecological Society of Australia 26, 231-239.

Mueller-Dombois D. (1973) A non-adapted vegetation interfered with water removal in a tropical rain forest area in Hawaii. Tropical Ecology 14, 1-18.

Noble J. C. (1989) Fire studies in mallee (Eucalyptus spp.) communities of western New South Wales: the effects of fires applied in different seasons on herbage productivity and their implications for management. Australian Journal of Ecology 14, 169-187.

203

Noble J. C. (1997) cited in Noble J. C. and Grice A. C. (2002), Fire regimes in semi-arid and tropical pastoral lands: managing biological diversity and ecosystem function. In: Flammable Australia. The Fire Regimes and Biodiversity of a Continent (eds R. A. Bradstock, R. A. Williams and M. A. Gill) pp. 373-400. Cambridge University Press, Cambridge, New York.

Noble J. C. and Grice A. C. (2002) Fire regimes in semi-arid and tropical pastoral lands: managing biological diversity and ecosystem function. In: Flammable Australia. The Fire Regimes and Biodiversity of a Continent (eds R. A. Bradstock, R. A. Williams and M. A. Gill) pp. 373-400. Cambridge University Press, Cambridge, New York.

Noble J. C., Grice A. C., Dobbie M. J., Muller W. J. and Wood J. T. (2001) Integrated shrub management in semi-arid woodlands of eastern Australia: effects of chemical defoliants applied after an initial disturbance. Rangeland Journal 23, 224-258.

Norman M. J. T. (1963) The pattern of dry matter and nutrient content changes in native pastures at Katherine, N. T. Australian Journal of Experimental Agriculture and Animal Husbandry 3, 119-124.

Norman M. J. T. (1969) The effect of burning and seasonal rainfall on native pasture at Katherine, N.T. Australian Journal of Experimental Agriculture and Animal Husbandry 9, 295-298.

Nurdin and Fulbright T. E. (1990) Germination of 2 legumes in leachate from introduced grasses. Journal of Range Management 43, 466-467.

Nye P. H. (1959) cited in Raison R. J. (1979), Modification of the soil environment by vegetaton fires, with particular reference to nitrogen transformations: a review. Plant and Soil 51, 73-108.

O'Connor K. F. and Powell A. J. (1963) cited in Daubenmire R. (1968), Ecology of fire in grasslands. Advances in Ecological Research 5, 209-266.

O'Connor T. G. (1991) Local extinction in perennial grasslands: a life-history approach. The American Naturalist 137, 753-773.

Ogle C. C., La Cock G. D., Arnold G. and Mickleson N. (2000) Impact of an exotic vine Clematis vitalba (F. Ranunculaceae) and the control measures on plant diversity in indigenous forest, Taihape, New Zealand. Austral Ecology 25, 539-551.

Olson B. E. (1999) Impacts of noxious weeds on ecological and economic systems. In: Biology and Management of Noxious Rangeland Weeds (eds R. L. Sheley and J. K. Petroff) pp. 4-18. Oregon State University Press, Corvallis.

Oram R. N. (1990) Register of Australian Herbage Plant Cultivars. 3rd edition. CSIRO Australian Herbage Plant Registration Authority, Division of Plant Industry, Canberra, A.C.T.

Orr D. M. (1998) A life cycle approach to the population ecology of two tropical grasses in Queensland, Australia. In: Population Biology of Grasses (ed. G. P. Cheplick) pp. 366-389. Cambridge University Press, Cambridge, United Kingdom.

Orr D. M. and Evenson C. J. (1993) A glasshouse study of competition between Astrebla lappacea (curly Mitchell grass) and two associated species. Rangeland Journal 15, 298-301.

Orr D. M. and Holmes W. E. (1984) Mitchell Grasslands. In: Management of Australia's

204

Rangelands (eds G. N. Harrington, A. D. Wilson and M. D. Young) pp. 241-254. CSIRO, Melbourne, Australia.

Orr D. M., McKeon G. M. and Day K. A. (1991) Burning and exclosure can rehabilitate degraded black speargrass (Heteropogon contortus) pastures. Tropical Grasslands 25, 333-336.

Orr D. M. and Paton C. J. (1997) Using fire to manage species composition in Heteropogon contortus (black speargrass) pastures. 2. Enhancing the effects of fire with grazing management. Australian Journal of Agricultural Research 48, 803-810.

Orr D. M., Paton C. J. and Blight G. W. (1996) An improved method for measuring the germinable soil seed banks of tropical pastures. Tropical Grasslands 30, 201-205.

Panetta F. D. and Lane D. W. (1996) Managing weeds in Australia's public lands and forests. In: Australian Weeds Conference: Proceedings Eleventh Australian Weeds Conference, Melbourne, Australia, 30 September-3 October, 1996 (ed. R. C. H. Shepherd) pp. 391-397. Weed Science Society of Victoria Inc, Frankston, Victoria.

Panetta F. D. and Scott J. K. (1995) Weeds and feral animals: management of future threats. In: Conserving Biodiversity: Threats and Solutions (eds R. A. Bradstock, T. D. Auld, D. A. Keith, R. T. Kingsford, D. Lunney and D. P. Sivertsen) pp. 220-228. Surrey Beatty and Sons, Chipping Norton, NSW, Australia.

Parker I. M., Simberloff D., Lonsdale W. M., Goodell K., Wonham M., Kareiva P. M., Williamson M. H., Von Holle B., Moyle P. B., Byers J. E. and Goldwasser L. (1999) Impact: toward a framework for understanding the ecological effects of invaders. Biological Invasions 1, 3-19.

Parker V. T. and Kelly V. R. (1989) Seed banks in California chaparral and other mediterranean climate shrublands. In: Ecology of Soil Seed Banks (eds M. A. Leck, V. T. Parker and R. L. Simpson) pp. 231-255. Academic Press, Inc., San Diego, California.

Parsons D. J. and Stohlgren T. J. (1989) Effects of varying fire regimes on annual grasslands in the southern Sierra Nevada of California. Madrono 36, 154-168.

Parsons J. J. (1972) Spread of African pasture grasses to the American tropics. Journal of Range Management 25, 12-17.

Parsons W. T. and Cuthbertson E. G. (2001) Noxious Weeds. CSIRO Publishing, Melbourne.

Pastor J., Downing A. and Erickson H. E. (1996) Species-area curves and diversity-productivity relationships in beaver meadows of Voyageurs National Park, Minnesota, USA. OIKOS 77, 399-406.

Paull C. J. and Lee G. R. (1978) Buffel grass in Queensland. Queensland Agricultural Journal 104, 57-75.

Peart D. R. (1979) Experiments on the biological significance of the morphology of seed-dispersal units in grasses. Journal of Ecology 67, 843-863.

Peart M. H. (1984) The effects of morphology, orientation and position of grass diaspores on seedling survival. Journal of Ecology 72, 437-453.

Pitts B. and Albrecht D. (2000) Buffel grass (Cenchrus ciliaris) control in central Australia. Danthonia 9, 7-8.

205

Pivello V. R., Carvalho V. M. C., Lopes P. F., Peccinini A. A. and Rosso S. (1999) Abundance and distribution of native and alien grasses in a "Cerrado" (Brazilian savanna) biological reserve. Biotropica 31, 71-81.

Pressland A. J. and Graham T. W. G. (1989) Approaches to the restoration of rangelands - the Queensland experience. Australian Rangeland Journal 11, 101-109.

Purdie R. W. and Slatyer R. O. (1976) Vegetation succession after fire in sclerophyll woodland communities in south-eastern Australia. Australian Journal of Ecology 1, 223-236.

QDE (1998a) Dalrymple National Park. Draft Management Plan. The State of Queensland. Department of Environment, Brisbane.

QDE (1998b) Moorrinya National Park. Draft Management Plan. State of Queensland. Department of Environment, Brisbane.

QPWS (2001) Effects of buffel grass pastures on native wildlife. Brochure produced by the Queensland Parks and Wildlife Service, Emerald Office.

Quirk M. F., Ash A. J. and McKillop G. (1997) Dalrymple Shire, QLD. Case study-the present. In: Sustainable Habitation in the Rangelands: Proceedings of a Fenner Conference on the Environment (eds N. O. J. Abel and S. Ryan) pp. 71-86. CSIRO Wildlife, Lyneham, A.C.T.

Rabinowitz D. and Rapp J. K. (1980) Seed rain in a North American tall grass prairie. Journal of Applied Ecology 17, 793-802.

Raison R. J. (1979) Modification of the soil environment by vegetation fires, with particular reference to nitrogen transformations: a review. Plant and Soil 51, 73-108.

Raison R. J. (1980) A review of the role of fire in nutrient cycling in Australian native forests, and of methodology for studying the fire-nutrient interaction. Australian Journal of Ecology 5, 15-21.

Raison R. J., Khanna P. K. and Woods P. V. (1985) Mechanisms of element transfer to the atmosphere during vegetation fires. Canadian Journal of Forestry Research 15, 132-140.

Ramakrishnan P. S. (1991 ) Ecology of biological invasion in the tropics: an overview. In: Ecology of Biological Invasion in the Tropics. (ed. P. S. Ramakrishnan) pp. 1-19. International Scientific Publications for National Institute of Ecology, New Delhi, India.

Ramakrishnan P. S. and Vitousek P. M. (1989) Ecosystem-level processes and the consequences of biological invasions. In: Biological Invasions: A Global Perspective (eds J. A. Drake and H. A. Mooney) pp. 218-300. Published on behalf of the Scientific Committee on Problems of the Environment of the International Council of Scientific Unions by Wiley, Chichester, New York.

Read T. R., Bellairs S. M., Mulligan D. R. and Lamb D. (2000) Smoke and heat effects on soil seed bank germination for the re-establishment of a native forest community in New South Wales. Austral Ecology 25, 48-57.

Rice E. L. (1972) Allelopathic effects of Andropogon virginicus and its persistence in old fields. American Journal of Botany 59, 752-754.

Rice E. L. (1979) Allelopathy-an update. The Botanical Review 45, 15-109.

206

Rice K. J. (1989) Impacts of seed banks on grassland community structure and population dynamics. In: Ecology of Soil Seed Banks (eds M. A. Leck, V. T. Parker and R. L. Simpson) pp. 211-230. Academic Press, Inc., San Diego, California.

Richardson D. M., Pysek P., Rejmanek M., Barbour M. G., Panetta F. D. and West C. J. (2000) Naturalization and invasion of alien plants: concepts and definitions. Diversity and Distributions 6, 93-107.

Rivas M. (1946) cited in Parsons J. J. (1972), Spread of African pasture grasses to the American tropics. Journal of Range Management 25, 12-17.

Roberts H. A. (1981) Seed banks in soils. Advances in Applied Biology 6, 1-55.

Robertson D. C., Morgan J. W. and White M. (1999) The use of prescribed fire to enhance control of English broom (Cytisus scoparius) invading a subalpine snowgum woodland in Victoria. Plant Protection Quarterly 14, 51-56.

Robin J. M. and Carr G. W. (1986) Hybridisation between introduced and native plants in Victoria, Australia. In: Ecology of Biological Invasions: An Australian Perspective (eds R. H. Groves and J. J. Burden) pp. Poster paper. Australian Academy of Science, Canberra.

Roche S., Dixon K. W. and Pate J. S. (1998) For everything a season: smoke-induced seed germination and seedling recruitment in a Western Australian Banksia woodland. Australian Journal of Ecology 23, 111-120.

Rogers L. G., Cannon M. G. and Barry E. V. (1999) Land Resources of the Dalrymple Shire. Department of Natural Resources, Brisbane, Qld.

Rosenzweig M. L. (1995) Species Diversity in Space and Time. Cambridge University Press, Cambridge, Britain.

Rossiter N. A., Setterfield S. A., Douglas M. M. and Hutley L. B. (2003) Testing the grass-fire cycles: alien grass invasion in the tropical savannas of northern Australia. Diversity and Distribution 9, 169-176.

Rothermel R. C. (1972) cited in Gill A. M. and Knight I. K. (1991), Fire measurement. In: Conference on Bushfire Modeling and Fire Danger Rating Systems. Proceedings, pp. 137-146. CSIRO, Canberra, ACT.

Russell-Smith J., Allan, G., Thackway, R., Roslying, T. and Smith, R. (2000) Fire management and savanna landscapes in northern Australia. In: Fire and Sustainable Agricultural and Forestry Development in Eastern Indonesia and Northern Australia. Proceedings of an International Workshop Held at Northern Territory Univerity, Darwin, Australia, 13-15 April 1999, (eds J. Russell-Smith, G. Hill, S. Djoeroemana, and B. Myres) The Australian Centre for International Agricultural Research, Canberra.

Sattler P. S. and Williams R. D. (1999) The Conservation Status of Queensland's Bioregional Ecosystems. Environmental Protection Agency, Brisbane.

Saucedo-Monarque E., Garcia-Moya E., Castellanos-Villegas A. E. and Flores-Flores J. L. (1997) Effect on species richness of the introduction of Buffel grass. Agrociencia 31, 83-90.

Scanlan J. C. (1980) Effects of spring wildfires on Astrebla (Mitchell grass) grasslands in north-west Queensland under varying levels of growing season rainfall. Australian Rangeland Journal 2, 162-168.

207

Scanlan J. C. and O'Rourke P. K. (1982) Effect of spring wildfires on Iseilema (Flinders grass) populations in the Mitchell grass region of north-western Queensland. Australia Journal of Botany 30, 591-600.

Schott G. W. and Hamburg S. P. (1997) The seed rain and seed bank of an adjacent native tallgrass prairie and old field. Canadian Journal of Botany 75, 1-7.

Schupp E. W. (1995) Seed-seedling conflicts, habitat choice, and patterns of plant recruitment. American Journal of Botany 82, 399-409.

Schwart M. W. (1997) Defining indigenous species: an introduction. In: Assessment and Management of Plant Invasions (eds J. O. Luken and J. W. Thieret) pp. 7-17. Springer-Verlag, New York.

Sharif-Zadeh F. and Murdoch A. J. (2000) The effects of different maturation conditions on seed dormancy and germination of Cenchrus ciliaris. Seed Science Research 10, 447-457.

Sharp D. and Simon B. K. (2002) Ausgrass: Grasses of Australia CD_ROM, Version 1.0 . Australian Biological Resources Study, Canberra, and Environmental Protection Agency, Quuensland.

Shaw N. H. (1957) Bunch spear grass dominance in burnt pastures in south-eastern Queensland. Australian Journal of Agricultural Research 8, 325-34.

Shea S. R., McCormick J. and Portlock C. C. (1979) The effects of fires on regeneration of leguminous species in the northern jarrah (Eucalyptus marginata Sm.) forest of western Australia. Australian Journal of Ecology 4, 195-205.

Silcock R. G. and Smith F. T. (1990) Viable seed retention under field conditions by western Queensland pasture species. Tropical Grasslands 24, 65-74.

Simpson R. L., Leck M. A. and Parker V. T. (1989) Seed banks: general concepts and methodological issues. In: Ecology of Soil Seed Banks (eds M. A. Leck, V. T. Parker and R. L. Simpson) pp. 3-8. Academic Press, Inc., San Diego, California.

Sindel B. M. (2000) Weeds and their impact. In: Australian Weed Management Systems (ed. B. M. Sindel) pp. 3-16. R.G. and F.J. Richardson, Meredith, Victoria.

Smith C. W. (1985) Impact of alien plants on Hawai'i's native biota. In: Hawai'i's Terrestrial Ecosystems: Preservation and Management. Proceedings of a Symposium Held June 5-6, 1984 at Hawai'i Volcanoes National Park (eds C. P. Stone and J. M. Scott) pp. 180-250. Cooperative National Park Resources Studies Unit, University of Hawaii, Honolulu.

Smith E. L. (1960) Effects of burning and clipping at various times during the wet season on tropical tall grass range in northern Australia. Journal of Range Management 13, 197-203.

Smith J. M. B. (1994) The changing ecological impact of broom (Cytisus scoparius) at Barrington Tops, New South Wales. Plant Protection Quarterly 9, 6-11.

Smith M. D. and Knapp A. K. (1999) Exotic plant species in a C4-dominanted grassland: invasibility, disturbance, and community structure. Oecologia 120, 605-612.

Stafford Smith M. (1994) Sustainable production systems and natural resource management in the rangelands. Outlook 148-159.

State of the Environment Advisory Council (Australia) (1996) Australia, State of the

208

Environment 1996. CSIRO Publishing, Collingwood, Victoria, Australia.

Stock W. D. and Lewis O. A. M. (1986) Soil nitrogen and the role of fire as a mineralizing agent in a South African coastal fynbos ecosystem. Journal of Ecology 74, 317-328.

Stocker G. C. and Mott J. J. (1981) Fire in the tropical forests and woodlands of northern Australia. In: Fire and the Australian Biota (eds A. M. Gill, R. H. Grove and I. R. Noble) pp. 425-439. Australian Academy of Science, Canberra.

Stocker G. C. and Sturtz J. D. (1966) The use of fire to establish Townsville lucerne in the Northern Territory. Australian Journal of Experimental Agriculture and Animal Husbandry 6, 277-279.

Stohlgren T. J., Schell L. D. and Heavel B. V. ( 1999) How grazing and soil quality affect native and exotic plant diversity in Rocky Mountain grasslands. Ecological Applications 9, 45-64.

Stuwe J. and Parsons R F (1977) Themada triandra grasslands on the Basalt Plains, Victoria: floristics and management effects. Australian Journal of Ecology 2, 467-476.

Swarbrick J. T. and Skarratt D. B. (1994) The Bushweed 2 Database of Environmental Weeds in Australia. The University of Queensland Gatton College, Gatton, Queensland.

Tainton N. M. (1981) Veld and Pasture Management in South Africa. Shuer and Shooter in association with the University of Natal Press, Natal.

Thomas L. K. (1980) cited in Woods K. D. (1997), Community response to plant invasion. In: Assessment and Management of Plant Invasions (eds J. O. Luken and J. W. Thieret) pp. 56-68. Springer-Verlag, New York.

Thompson K. (1992) The functional ecology of seed banks. In: Seeds. The Ecology of Regeneration in Plant Communities (ed. M. Fenner) pp. 231-258. CAB International, Wallingford, Oxon U.K.

Thompson L. and Harper J. L. (1988) The effect of grasses on the quality of transmitted radiation and its influence on the growth of white clover Trifolium repens. Oecologia 75, 343-347.

Tolhurst K. (1996) Effects of fuel reduction burning on flora in a dry sclerophyll forest. In: Fire and Biodiversity. The Effects and Effectiveness of Fire Management. Proceedings of the Conference Held 8-9 October, Footscray, Melbourne., Department of the Environment, Sport and Territories, Biodiversity Unit, Canberra, ACT.

Tothill J. C. (1969) Soil temperature and seed burial in relation to the performance of Heteropogon contortus and Themeda triandra in burnt native woodland pastures in eastern Queensland. Australian Journal of Botany 17, 269-275.

Tothill J. C. (1971) A review of fire in the management of native pasture with particular reference to north-eastern Australia. Tropical Grasslands 5, 1-10.

Tothill J. C. (1977) Seed germination studies with Heteropogon contortus. Australian Journal of Ecology 2, 477-484.

Tothill J. C. and Hacker J. B. (1983) The Grasses of South East Queensland. University of Queensland Press, St Lucia, Queensland.

209

Tothill J. C., Hargreaves J. N. G., Jones R. M. and McDonald C. K. (1992) BOTANAL - A Comprehensive Sampling and Computing Procedure for Estimating Pasture Yield and Composition. 1 Field Sampling. Tropical Agronomy Technical Memorandum No. 78. CSIRO Division of Tropical Crops and Pastures, Brisbane.

Townsend S. A. and Douglas M. M. (2000) The effect of three fire regimes on stream water quality, water yield and export coefficients in a tropical savanna (northern Australia). Journal of Hydrology 229, 118-137.

Trollope W. S. W. (1997) Behaviour, effects and use of fire in the savanna of southern Africa. In: Fire in the Management of Northern Australian Pastoral Lands pp. 9-23. The Tropical Grassland Society of Australia, Brisbane, Australia.

Turner C. L., Blair J. M., Schartz R. J. and Neel J. C. (1997) Soil N and plant responses to fire, topography, and supplemental N in tallgrass prairie. Ecology 78, 1832-1843.

Tyler C. M. (1994) cited in Tyler C. M. (1995), Factors contributing to postfire seedling establishment in chaparral: direct and indirect effects of fire. Journal of Ecology 83, 1009-1020.

Tyler C. M. (1995) Factors contributing to postfire seedling establishment in chaparral: direct and indirect effects of fire. Journal of Ecology 83, 1009-1020.

Usher M. B. (1986) Invasibility and wildlife conservation: invasive species on nature reserves. Philosophical Transactions of the Royal Society of London 314, 695-710.

Valone T. J. and Kelt D. A. (1999) Fire and grazing in a shrub-invaded arid grassland community: independent or interactive ecological effects? Journal of Arid Environments 42, 15-28.

Van de Vijver C. A. D. M., Poot P. and Prind H. H. T. (1999) Causes of increased nutrient concentrations in post-fire regrowth in an East African savanna. Plant and Soil 214, 173-185.

van Wilgen B. W. and Richardson D. M. (1985) The effects of alien shrub invasion on vegetation structure and fire behaviour in South African fynbos shrublands: a simulation study. Journal of Applied Ecology 22, 955-966.

Vermeij G. J. (1996) An agenda for invasion biology. Biological Conservation 78, 3-9.

Versfeld D. B. and van Wilgen B. W. (1986) cited in Walker L. R. and Smith S. D. (1997), Impacts of invasive plants on community and ecosystem properties. In: Assessment and Management of Plant Invasions (eds J. O. Luken and J. W. Thieret) pp. 69-86. Springer-Verlag, New York.

Vitelli J.S. (2000) Options for effective weed management. In: Pastures for Production and Protection. Tropical Grasslands Conference, Emerald 26-28 April 2000, Tropical Grassland Society of Australia, St Lucia, Queensland.

Vitousek P. M. (1986) Biological invasions and ecosystem properties: can species make a difference? In: Ecology of Biological Invasions of North America and Hawaii (eds H. A. Mooney and J. A. Drake) pp. 163-176. Springer, New York .

Vitousek P. M. (1990) Biological invasions and ecosystem processes: towards an integration of population biology and ecosystem studies. OIKOS 57, 7-13.

210

Vitousek P. M., D'Antonio C. M., Loope L. L., Rejmánek M. and Westbrooks R. (1997) Introduced species: a significant component of human-caused global change. New Zealand Journal of Ecology 21, 1-16.

Vitousek P. M. and Walker L. R. (1989) Biological invasions by Myrica faya in Hawai'i: plant demography, nitrogen fixation, ecosystem effects. Ecological Monographs 59, 247-265.

Vitousek P. M., Walker L R., Whiteaker L D., Mueller-Dombois D. and Matson P A. (1987) Biological invasion by Myrica faya alters ecosystem development in Hawaii. Science 238, 802-4.

Vlok J. H. J. (1988) Alpha diversity of lowland fynbos herbs at various levels of infestation by alien annuals. African Journal of Botany 54, 623-627.

Vogl R. J. (1974) Effects of fire on grasslands. In: Fire and Ecosystems (eds T. T. Kozlowski and C. E. Ahlgren) pp. 139-94. Academic Press, New York.

Vogl R. J. (1975) Fire: a destructive menace or natural process. In: Recovery and Restoration of Damaged Ecosystems (eds J. Cairns Jr., K. L. Dickson and E. E. Herricks) pp. 261-289. University Press of Virginia, Charlottsville, Virginia, USA.

Vogl R. J. (1977) cited in Christensen P. E. and Burrows N. D. (1986), Fire: an old tool with a new use. In: Ecology of Biological Invasions: an Australian Perspective (eds R. H. Groves and J. J. Burden) pp. 97-105. Australian Academy of Science, Canberra.

Vranjic J. A., Woods M. J. and Barnard J. (2000) Soil-mediated effects on germination and seedling growth of coastal wattle (Acacia sophprae) by the environmental weed, bitou bush (Chryanthemoides monilifera ssp. rotundata). Austral Ecology 25, 445-453.

Wagner W.H. (1993) Problems with biotic invasives: a biological viewpoint. In: Biological Pollution: The Control and Impact of Invasive Exotic Species. Proceedings of a Symposium Held at Indianapolis, Indiana, USA, 25-26 October 1991, (ed. B. N. McKnight) pp. 1-8. Indiana Academy of Science, Indianapolis, USA, India.

Wagner W. L., Herbst D. R. and Sohmer S. H. (1990) Manual of the Flowering Plants of Hawaii. University of Hawaii Press, Honolulu.

Waide R. B., Willig M. R., Steiner C. F., Mittelbach G., Gough L., Dodson S. I., Juday G. P. and Parmenter R. (1999) The relationship between productivity and species richness. Annual Review of Ecology and Systematics 30, 257-300.

Walker J. (1981) Fuel dynamics in Australian vegetation. In: Fire and the Australian Biota (eds A. M. Gill, R. H. Grove and I. R. Noble) pp. 101-127. Australian Academy of Science, Canberra.

Walker J., Condon R. W., Hodgkinson K. C. and Harrington G. N. (1981) Fire in pastoral areas of poplar box (Eucalyptus populnea) lands. Australian Rangeland Journal 3, 12-23.

Walker J., Malafant, K. and Edwards, J. (1983) The effects of fire on a Themeda-Heteropogon grassland near Rockhampton. In: The Second Queensland Fire Research Workshop, July 5 to July 7 1983, Held at the Forestry Training Centre, Gympie, (ed. B. B. Roberts) pp. 85-94. Darling Downs Institute of Advanced Education, Toowoomba.

Walker J. and Peet R. K. (1983) Composition and species diversity of pine-wiregrass savannas of the Green Swamp, North Carolina. Vegetatio 55, 163-179.

211

Walker L. R. and Smith S. D. (1997) Impacts of invasive plants on community and ecosystem properties. In: Assessment and Management of Plant Invasions (eds J. O. Luken and J. W. Thieret) pp. 69-86. Springer-Verlag, New York.

Walsh R. P. D. and Voigt P. J. (1977) Vegetation litter: an underestimated variable in hygrology and geomorphology. Journal of Biogeography 4, 253-274.

Wapshere A. J. (1990) Biological control of grass weeds in Australia: an appraisal. Plant Protection Quarterly 5, 62-75.

Warr S. J., Thompson K. and Kent M. (1993) Seed banks as a neglected area of biographic research: a review of literature and sampling techniques. Progress in Physical Geography 17, 329-347.

Watson S. (1995) cited in Lunt I. D. and Morgan J. W. (2002), The role of fire regimes in temperate lowland grasslands of southeastern Australia. In: Flammable Australia. The Fire Regimes and Biodiversity of a Continent (eds R. A. Bradstock, R. A. Williams and M. A. Gill) pp. 178-196. Cambridge University Press, Cambridge, New York.

Wedin D. A. and Tilman D. (1990) Species effects on nitrogen cycling: a test with perennial grasses. Oecologia 84, 433-441.

Weiher E. (1999) The combined effects of scale and productivity on species richness. Journal of Ecology 87, 1005-1011.

Wells M. J., Poynton R. J., Balsinhas A. A., Musil K. J., Joffe H., van Hoepen E. and Abbott S. K. (1986) cited in Fox M. D. (1995), Conserving biodiversity: impact and management of exotic organisms. In: Conserving Biodiversity: Threats and Solutions (eds R. A. Bradstock, T. D. Auld, D. A. Keith, R. T. Kingsford, D. Lunney and D. P. Sivertsen) pp. 177-183. Surrey Beatty and Sons, Chipping Norton, NSW, Australia.

West O. (1965) Fire in Vegetation and Its Use in Pasture Management, With Special Reference to Tropical and Subtropical Africa. Commonwealth Bureau of Pastures and Field Crops, Hurley, Berkshire.

Wheeler B. D. and Giller K. E. (1982) Species richness of herbaceous fen vegetation in Broadland, Norfolk in relation to the quantity of above-ground plant material. Journal of Ecology 70, 179-200.

Whelan R. J. (1995) The Ecology of Fire. Cambridge University Press, Cambridge, United Kingdom.

Whisenant S. G. (1990a) cited in DiTomaso J. M. (2000), Invasive weeds in rangelands: species, impacts and management. Weed Science 48, 255-265.

Whisenant S. G. (1990b) Postfire population dynamics of Bromus japonicus. American Midland Naturalist 123, 301-308.

White (1997) From the guest editor. Alice Springs Rural Review 47, 2.

Whitehead P. and Wilson, C. (2000) Exotic grasses in northern Australia: species that should be sent home. In: Northern Grassy Landscapes Conference, Katherine, Northern Territory 29-31 August 2000, Cooperative Reseach for the Sustainable Development of Tropical Savannas, Darwin, Northern Territory.

Whitford W. G., Nielson R. and de Soyza A. (2001) Establishment and effects of

212

establishment of creosotebush, Larea tridentata, on the Chihuahuan desert watershed. Journal of Arid Environments 47, 1-10.

Williams D. and Cook G. (2001) Savanna fire regimes. In: Savanna Burning. Understanding and Using Fire in Northern Australia (eds R. Dyer, P. Jacklyn, I. Partridge, J. Russell-Smith and D. Williams) pp. 15-28. Tropical Savanna CRC, Darwin.

Williams D. G. and Baruch Z. (2000) African grass invasion in the Americas: ecosystem consequences and the role of ecophysiology. Biological Invasions 2, 123-140.

Williams J. A. and West C. J. (2000) Environmental weeds in Australia and New Zealand: issues and approaches to management. Austral Ecology 25 , 425-444.

Williams J. R. (1995) Tree mortality in relation to fire intensity in a tropical savanna of the Kakadu region, Northern Territory, Australia. CALM Science Supplement 4, 77-82.

Williams M. R. (1995) An extreme-value function model of the species incidence and species-area relations. Ecology 76, 2607-2616.

Williams M. R. (1996) Species-area curves: the need to include zeroes. Global Ecology and Biogeography Letters 5, 91-93.

Williams P. R. (2000) Fire-simulated rainforest seedling recruitment and vegetative regeneration in a densely grassed wet sclerophyll forest of north-eastern Australia. Australian Journal of Botany 48, 651-658.

Williams P. R., Congdon R. A., Grice A. C. and Clarke P. J. (2003a) Fire-related cues break seed dormancy of six legumes of tropical eucalypt savannas in north-eastern Australia. Austral Ecology 28, 507-541.

Williams P. R., Congdon R. A., Grice A. C. and Clarke P. J. (2003b) Effects of fire regime on plant abundance in a tropical eucalypt savanna of north-eastern Australia. Austral Ecology 28, 327-338.

Williams R. J., Cook G. D., Lugwig J. A. and Tongway D. (1997) Torch, trees, teeth and tussocks: disturbances in the tropical savanna of the Northern Territory. In: Frontiers in Ecology. Building the Links (eds N. Klomp and I. Lunt) pp. 55-67. Elsevier Science, Oxford.

Williams R. J., Griffiths A. D. and Allan G. E. (2002) Fire regimes and biodiversity in the savannas of northern Australia. In: Flammable Australia. The Fire Regimes and Biodiveristy of a Continent (eds R. A. Bradstock, R. A. Williams and M. A. Gill) pp. 281-304. Cambridge University Press, Cambridge, New York.

Williamson M. (1996) Biological Invasions. Chapman & Hall, London.

Wilson C. and Mudita, W. (2000) Fire and Weeds: Interactions and management implications. In: Fire and Sustainable Agricultural and Forestry Development in Eastern Indonesia and Northern Australia. Proceedings of an International Workshop Held at Northern Territory University, Darwin, Australia, 13-15 April 1999, (eds J. Russell-Smith, G. Hill, S. Djoeroemana, and B. Myres) The Australian Centre for International Agricultural Research, Canberra.

Winkel V. K., Roundy B. A. and Cox J. R. (1991) Influence of seedbed microsite characteristics on grass seedling emergence. Journal of Range Management 44, 210-214.

Winkworth R. E. (1971) Longevity of buffel grass seed sown in an arid Australian range.

213

Journal of Range Management 24, 141-145.

Wisheu I. C. and Keddy P. A. (1989) Species richness - standing crop relationships along four lakeshore gradients: constraints on the general model. Canadian Journal of Botany 67, 1609-1617.

Witkowski E. T. F. and Mitchell D. T. (1987) Variations in soil phosphorus in the fynbos biome, South Africa. Journal of Ecology 75, 1159-1171.

Woinarski J. (2001) A review of changes in status and threatening processes. In: Developing an Analytical Framework for Monitoring Biodiversity in Australia's Rangelands. A Report by the Tropical Savanna CRC for the National Land and Water Resources Audit (eds P. Whitehead, J. Woinarski , A. Fisher, R. Fensham and K. Beggs) pp. 71-310. Tropical Savanna CRC, Darwin.

Woods K. D. (1997) Community response to plant invasion. In: Assessment and Management of Plant Invasions (eds J. O. Luken and J. W. Thieret) pp. 56-68. Springer-Verlag, New York.

Wright H. A. and Klemmedson J. O. (1965) Effect of fire on bunchgrasses of the sagebrush-grass region in southern Idaho. Ecology 46, 680-688.

Xiong S. and Nilsson C. (1999) The effects of plant litter on vegetation: a meta-analysis. Journal of Ecology 87, 984-994.

Young J. A., Evans R. A., Raguse C. A. and Larson J. R. (1981) Germinable seeds and periodicity of germination in annual grasslands. Hilgardia 49, 1-37.

Zacharias P. J. K., Tainton N. M. and Oberholster C. (1988) The effect of fire on germination in five common veld grasses. Journal of the Grassland Society of South Africa 5, 229-230.

214

APPENDICES

Appendix 1A. Herbaceous species found in C. ciliaris and non-C. ciliaris plots in the

Dalrymple Shire survey. Species classified as rare, that is found in only one plot in the

survey, are denoted by *.

FORBS Status Acanthospermum hispidum DC. exotic * Achyranthes aspera L. exotic * Alternanthera sp. unknown Blumea sp. unknown Boerhavia schomburgkiana Oliv. native Brunoniella acaulis (R.Br.) Bremek. native Camptacra barbata N.T.Burb. native Chamaesyce hirta (L.) Millsp. exotic Chamaesyce mitchelliana (Boiss.) D.C.Hassall native Chenopodium carinatum R.Br. native * Commelina cyanea R.Br. native * Corchorus sp. native Cucumus sp. native Drosera sp. native * Epaltes australis Less. native * Evolvulus alsinoides L. native Goodenia glabra R.Br. native Goodenia sp. native Grewia retusifolia Kurz native Heliotropium pauciflorum R.Br. native Hibiscus sp. native * Hybanthus enneaspermus (L.) F.Muell. native Ipomoea coptica (L.) Roth ex Roem. & Schult. native Ipomoea eriocarpa R.Br. native Ipomoea polymorpha Roem. & Schult. native Marsdenia viridiflora R.Br. native Melhania oblongifolia F.Muell. native Mitrasacme pygmaea R.Br. native * Ocimum sp. unknown Oldenlandia mitrasacmoides (F.Muell.) F.Muell. exotic Peripleura hispidula (F.Muell. ex A.Gray) G.L.Nesom native * Phyllanthus virgatus G.Forst. native Polycarpaea corymbosa (L.) Lam. native * Polygala linariifolia Willd. native Polymeria calycina R.Br. native Portulaca oleracea L. native Pterocaulon redolens (Willd.) Fern.-Vill. native * Rostellularia adscendens (R.Br.) R.M.Barker native Sida fibulifera Lindl. native Sida subspicata F.Muell. ex Benth. native Sida trichopoda F.Muell. native Solanum sp. unknown Spermacoce brachystema R.Br. ex Benth. native Thysanotus sp. native *

215

Trianthema triquetra Rottb. Ex Willd. native * Wahlenbergia sp. native * Waltheria indica L. native LEGUMES Alysicarpus sp. unknown *Cajanus scarabeoides (L.) Thouars native Chamaecrista absus (L.) H.S.Irwin & Barneby var. absus native Crotalaria medicaginea Lam. native Crotalaria montana Roth native Crotalaria verrucosa L. native * Desmodium sp. native Galactia tenuiflora (Spreng.) Willd. ex Wight & Arn. native Glycine tomentella Hayata native * Indigastrum parviflorum (B.Heyne ex Wight & Arn.) Schrire native * Indigofera colutea (Burm.f.) Merr. native Indigofera haplophylla F. Muell. native * Indigofera linifolia (L.f.) Retz. native Indigofera linnaei Ali native Macroptilium atropurpureum (DC.) Urb. exotic * Neptunia sp. native Rhynchosia minima (L.) DC. native Stylosanthes hamata (L.) Taub. exotic Stylosanthes humilis Kunth exotic Stylosanthes scabra Vogel exotic Tephrosia filipes Benth. native Tephrosia leptoclada Benth. native Tephrosia rosea F.Muell. ex Benth. native Vigna lanceolata Benth. native Zornia muellerana Mohlenbr. native Zornia muriculata Mohlenbr. native PERENNIAL GRASSES Alloteropsis cimicina (Retz.) Stapf native Aristida benthamii Henrard native Aristida calycina R.Br. native Aristida holathera Domin native Aristida ingrate Domin native Aristida jerichoensis (Domin) Henrard native Aristida perniciosa Domin native Bothriochloa decipiens (Hack.) C.E.Hubb. native * Bothriochloa ewartiana (Domin) C.E.Hubb. native Bothriochloa pertusa (L.) A.Camus exotic Cenchrus ciliaris L. exotic Chloris pectinata Benth. native Chloris virgata Sw. native Chrysopogon fallax S.T.Blake native Dichanthium sericeum (R.Br.) A.Camus native Digitaria ammophila (F.Muell.) Hughes native Enneapogon polyphyllus (Domin) N.T.Burb. native Eriachne sp. native Eriochloa sp. native Heteropogon contortus (L.) P.Beauv. ex Roem. & Schult. native Melinis repens (Willd.) Zizka exotic

216

Panicum effusum R.Br. native Paspalidium sp. native * Sehima nervosum (Rottler) Stapf native * Sarga (Sorghum) plumosum (R.Br.) P.Beauv. native Themeda triandra Forssk. native Tripogon loliiformis (F.Muell.) C.E.Hubb. native Urochloa mosambicensis (Hack.) Dandy exotic ANNUAL GRASSES Brachiaria windersii C.E.Hubb. native * Brachyachne convergens (F.Muell.) Stapf native Dactyloctenium radulans (R.Br.) P.Beauv. native Digitaria ciliaris (Retz.) Koeler exotic Echinochloa sp. unknown Eragrostis brownii (Kunth) Nees ex Steud. native Iseilema sp. native * Mnesithea formosa (R.Br.) de Koning & Sosef native Perotis rara R.Br. native * Schizachyrium fragile (R.Br.) A.Camus native Sporobolus australasicus Domin native Thaumastochloa pubescens (Benth.) C.E.Hubb. native * Tragus australianus S.T.Blake native SEDGES Cyperus bifax C.B.Clarke native Cyperus gilesii Benth. native Fimbristylis sp. native Scleria sp. native *

217

Appendix 1B. Herbaceous species found in surveyed plots at Hillgrove.

FORBS Indigofera colutea (Burm.f) Merr. Acanthospermum hispidum DC. Indigofera hirsuta L. Alternanthera sp. Indigofera linifolia (L.f.) Retz. Boerhavia schomburgkiana Oliv. Indigofera linnaei Ali Camptacra barbata N.T.Burb. Rhynchosia minima (L.) DC. Chamaesyce mitchelliana Boiss D.C.Hassall Stylosanthes hamata (L.) Taub. Corchorus trilocularis L. Stylosanthes scabra Vogel Euphorbia hirta L. Vigna lanceolata Benth. Evolvulus alsinoides L. Zornia dyctiocarpa DC. Grewia retusifolia Kurz Zornia muriculata Mohlenbr. Hybanthus enneaspermus (L.) F.Muell. Ipomoea eriocarpa R.Br. ANNUAL GRASSES Ipomoea gracilis R.Br. Brachyachne convergens (F.Muell.) Stapf Jacquemontia paniculata Digitaria ciliaris (Retz.) Koeler Marsdenia viridiflora R.Br. Eragrostis sororia Domin Melhania oblongifolia F.Muell. Oxalis corniculata L. PERENNIAL GRASSES Phyllanthus maderspatensis L. Aristida calycina R.Br. Polymeria sp. Aristida holathera Domin Portulaca filifolia F.Muell. Aristida leptopoda Benth. Pterocaulon redolens (Willd.) Fern.-Vill. Bothriochloa decipiens (Hack.) C.E.Hubb. Rostellularia adscendens R.M.Barker Bothriochloa ewartiana (Domin) C.E.Hubb. Sida acuta/rohlenae Bothriochloa pertusa (L.) A.Camus Sida fibulifera Lindl. Cenchrus ciliaris L. Sida spinosa L. Chrysopogon fallax S.T.Blake Dichanthium fecundum S.T.Blake LEGUMES Dichanthium sericeum (R.Br.) A.Camus Chamaecrista absus (L.) H.S.Irwin & Barneby Enneapogon polyphyllus (Domin) N.T.Burb. Crotalaria goreensis Guill. & Perr. Eriochloa procera (Retz.) C.E.Hubb. Crotalaria juncea L. Heteropogon contortus (L.) Crotalaria medicaginea Lam. Melinis repens (Willd.) Zizka Crotalaria montana Roth Panicum decompositum R.Br. Crotalaria verrucosa L. Tripogon loliiformis (F.Muell.) C.E.Hubb. Desmodium sp. Urochloa mosambicensis (Hack.) Dandy Galactia Muelleri Benth. Glycine tabacina (Labill.) Benth. SEDGES Glycine tomentella Hayata Cyperus sp.

218

Appendix 2A. Aussie peat components.

Aussie peat is an organic product made from boiling and milling pine bark and wood chip,

producing a product with texture identical to peat moss.

Chemical analyses:

Air filled porosity (%) 12

Water holding capacity (%) 72

pH approx 6.4

Conductivity (dS/m) 0.11

Appendix 2B. Germinable seed content of Cenchrus ciliaris and Heteropogon contortus

seed material.

To determine the amount of germinable seed in C. ciliaris and H. contortus seed material,

four samples of the diaspores of both species were weighed (0.34 g mean weight) and the

seeds from each sample were then separated from their appendages. Germination tests were

conducted using all the seed from each H. contortus sample (mean of 55 seeds/sample). For

each C. ciliaris sample, only 50 seeds/sample were used because of the high numbers of seed

found in each sample (mean of 117 seeds/sample). Germination was determined for each

sample by placing the seed on filter paper in covered petri dishes, one dish per sample. Seed

was kept moist and germination was monitored in ambient laboratory conditions over nine

days. The average germinability of seed was 10% and 19% for H. contortus and C ciliaris

respectively.

219

Appendix 3A. Herbaceous species found in the Dalrymple seed banks.

FORBS LEGUMES Abutilon calliphyllum Domin Chamaecrista absus (L.) H.S.Irwin & Barneby* Amaranthus cochleitepalus Domin Indigofera colutea (Burm.f.) Merr.* Ammania multiflora Roxb. Zornia muellerana Mohlenbr. Blumea saxitilis Zoll. & Moritzi Corchorus tridens L. ANNUAL GRASSES Dysphania glomulifera (Nees) Paul G.Wilson

Dactyloctenium radulans (R.Br.) P.Beauv.* Perotis rara R.Br.*

Epaltes australis Less. Tragus australianus S.T.Blake Glinus oppositifolius (L.) A.DC.* Heliotropium ovalifolium Forssk. PERENNIAL GRASSES Hybanthus enneaspermus (L.) F.Muell.* Bothriochloa pertusa (L.) A.Camus* Ipomoea polymorpha Roem. & Schult. Brachiaria subquadripara (Trin.) Hitchc. Melhania oblongifolia F.Muell.* Cenchrus ciliaris L.* Mitrasacme pygmaea R.Br. Enneapogon polyphyllus (Domin) N.T.Burb. Ocimum sp. Eragrostis brownii (Kunth) Nees ex Steud. Oldenlandia coerulescens (F.Muell.) F.Muell.

Eragrostis tenellula (Kunth) Steud. Heteropogon contortus (L.) P.Beauv. ex Roem.

Phyllanthus sp. Portulaca filifolia F.Muell.

& Schlt.* Melinis repens (Willd.) Zizka*

Portulaca oleracea L. Sporobolus australasicus Domin Pterocaulon redolens (Willd.) Fern.-Vill. Salsola kali L. SEDGES Scoparia dulcis L. Bulbostylus barbata (Rottb.) C.B.Clarke Sida spinosa L. Cyperus difformis L. Tribulus sp. Cyperus squarrosus L. Wahlenbergia caryophylloides P.J.Sm. Fimbristylis bisumbellata (Forssk.) Bubani * indicates species also found (and identified to species level) in the standing vegetation.

220

Appendix 3B. Herbaceous species found in the Moorrinya seed banks.

FORBS LEGUMES Abutilon malvifolium (Benth.) J.M.Black* Aeschynomene indica L. Alternanthera nodiflora R.Br. Ammania multiflora Roxb. ANNUAL GRASSES Bergia trimera Fisch. & C.A.Mey. Brachyachne sp. Blumea diffusa R.Br. ex Benth.* Dactyloctenium radulans (R.Br.) P.Beauv.* Centipeda minima (L.) A.Braun & Asch. Iseilema vaginiflorum Domin Chamaesyce hirta (Kunth) Steud. Dysphania glomulifera (Nees) Paul G.Wilson PERENNIAL GRASSES Epaltes australis Less. Cenchrus ciliaris L.* Ipomoea lonchophylla J.M.Black Dichanthium sericeum (R.Br.) A.Camus* Murdannia graminea (R.Br.) G.Brueckn. Enneapogon avenaceus (Lindl.) C.E.Hubb. Oldenlandia coerulescens (F.Muell.) F.Muell.

Eragrostis kennedyae F.Turner Eragrostis tenellula (Kunth) Steud.

Oldenlandia mitrasacmoides (F.Muell.) F.Muell.

Sporobolus australasicus Domin xxxx

Phyllanthus sp. SEDGES Portulaca oleracea L. Cyperus difformis L.* Sclerolaena ramulosa (C.T.White) A.J.Scott Cyperus gilesii Benth.* Sclerolaena bicornis Lindl.* Cyperus squarrosus L. Sida fibulifera Lindl.* Fimbristylis subartistata Benth. Sida spinosa L. Streptoglossa adscendens (Benth.) Dunlop Trianthema triquetra Rottb. ex Willd.*

* indicates species also found (and identified to species level) in the standing vegetation.

221

Appendix 4A. Herbaceous species found in burning treatment plots at Dalrymple.

FORBS Indigofera linnaei Ali Abutilon sp. Rhynchosia minima (L.) DC. Ageratum sp. Senna occidentalis (L.) Link Alternanthera sp. Stylosanthes sp. Boerhavia sp. Vigna sp. Brunoniella acaulis (R.Br.) Bremek. Zornia sp. Commelina cyanea R.Br. Corchorus sp. ANNUAL GRASSES Crinum sp. Dactyloctenium radulans (R.Br.) P.Beauv.* Emilia sonchifolia (L.) DC. Perotis rara R.Br.* Euphorbia sp. Sporobolus sp. Evolvulus alsinoides L. Glinus oppositifolius (L.) A.DC.* PERENNIAL GRASSES Hibiscus meraukensis Hochr. Aristida sp. Hybanthus enneaspermus (L.) F.Muell.* Bothriochloa decipiens (Hack.) C.E.Hubb. Ipomoea sp. Bothriochloa ewartiana (Domin) C.E.Hubb. Malvastrum americanum (L.) Torr. Bothriochloa pertusa (L.) A.Camus* Melhania oblongifolia F.Muell.* Cenchrus ciliaris L.* Ocimum sp. Chloris sp. Parsonsia sp. Chrysopogon fallax S.T.Blake Passiflora foetida L. Cymbopogon bombycinus (R.Br.) Domin Phyllanthus sp. Dichanthium sericeum (R.Br.) A.Camus Polygala sp. Digitaria sp. Portulaca sp. Enneapogon sp. Pseuderanthemum sp. Eragrostis sp. Pterocaulon sp. Rostellularia adscendens (R.Br.)

Heteropogon contortus (L.) P.Beauv. ex Roem. & Schlt.*

R.M.Barker Melinis repens (Willd.) Zizka* Sida sp. Panicum sp. Spermacoce brachystema R.Br. ex Benth. Tripogon loliiformis (F.Muell.) C.E.Hubb. Tricoryne anceps R.Br. Urochloa mosambicensis (Hack.) Dandy Vernonia sp. Wahlenbergia sp. SEDGES Waltheria indica L. Fimbristylis sp. LEGUMES Chamaecrista absus (L.) H.S.Irwin & Barneby*

Crotalaria medicaginea Lam. Crotalaria montana Roth Crotalaria novae-hollandiae DC. Crotalaria verrucosa L. Glycine tomentella Hayata Indigofera colutea (Burm.f.) Merr.* Indigofera hirsuta L. Indigofera linifolia (L.f.) Retz.

* indicates species also found (and identified to species level) in the soil seed bank.

222

Appendix 4B. Herbaceous species found in burning treatment plots at Moorrinya.

FORBS LEGUMES Abutilon malvifolium (Benth.) J.M.Black* Aeschynomene sp. Abutilon sp. Crotalaria brevis Domin Alternanthera sp. Crotalaria montana Roth Blumea diffusa R.Br. ex Benth.* Indigofera sp. Boerhavia sp. Neptunia sp. Commelina sp. Rhynchosia minima (L.) DC. Corchorus sp. Epaltes sp. ANNUAL GRASSES Euphorbia sp. Brachiaria sp. Evolvulus alsinoides L. Dactyloctenium radulans (R.Br.) P.Beauv.* Gomphrena celosioides Mart. Echinochloa sp. Hibiscus trionum L. Elytrophorus spicatus (Willd.) A.Camus Ipomoea sp. Iseilema sp. Malvastrum americanum (L.) Torr. Sporobolus australasicus Domin Marsilea sp. Melhania oblongifolia F.Muell. PERENNIAL GRASSES Minuria integerrima (DC.) Benth. Aristida sp. Ocimum sp. Astrebla lappacea (Lindl.) Domin Oldenlandia sp. Astrebla squarrosa C.E.Hubb. Peplidium foecundum W.R.Barker Bothriochloa ewartiana (Domin) C.E.Hubb. Phyllanthus sp. Cenchrus ciliaris L.* Polymeria sp. Chloris sp. Portulaca sp. Dichanthium sericeum (R.Br.) A.Camus Pterocaulon sp. Dichanthium setosum S.T.Blake* Ptilotus obovatus (Gaudich.) F.Muell. Digitaria sp. Salsola kali L. Enneapogon polyphyllus (Domin) N.T.Burb. Sclerolaena anisacanthoides (F.Muell.) Domin

Eragrostis sp. Heteropogon contortus (L.) P.Beauv. ex

Sclerolaena bicornis Lindl.* Roem. & Schlt. Sclerolaena muricata (Moq.) Domin Panicum sp. Sida acuta Burm.f. Sporobolus actinocladus (F.Muell.) F.Muell. Sida fibulifera Lindl.* Sporobolus caroli Mez Sida trichopoda F.Muell. Tripogon lolliformis (F.Muell.) C.E.Hubb. Solanum sp. Spermacoce sp. SEDGES Streptoglossa sp. Cyperus difformis L.* Trianthema triquetra Rottb. ex Willd.* Cyperus gilesii Benth.* Fimbristylis sp.

* indicates species also found (and identified to species level) in the soil seed bank.

223

IMPACTS AND MANAGEMENT OF CENCHRUS CILIARIS (BUFFEL … · 3.2.3. Fire as a management tool for...

Documents

Transcript of IMPACTS AND MANAGEMENT OF CENCHRUS CILIARIS (BUFFEL … · 3.2.3. Fire as a management tool for...